JP2018023478A - Fundus image processing device - Google Patents

Abstract

Provided is a fundus image processing apparatus capable of objectively measuring a blood flow width ratio while solving a problem of variation due to a subject's subjectivity and measurement location. A display means for displaying a measurement graphic, a measurement range setting means for setting two parallel sides of the measurement graphic as a measurement range based on the measurement graphic arranged on the fundus image, a measurement range Create a luminance array that is an array of luminance values corresponding to each sample obtained by sampling the sides set as, create a differential array that is an array of differential luminance values using the luminance array, The maximum sample having the maximum value and the minimum sample having the minimum value are searched, the second maximum sample is searched based on the maximum sample, the second minimum sample is searched based on the minimum sample, and the maximum sample and the minimum sample are searched. The blood flow region end point specifying means 50 for specifying the two end points of the blood flow region based on the second and the two boundary points of the blood column reflection region based on the second maximum sample and the second minimum sample. [Selection] Figure 2

Description

  The present invention relates to a fundus image processing technique, and more particularly to a technique for analyzing information on a blood flow width such as a blood column reflex width, which is one of indices of arteriosclerosis and hypertension, by processing the fundus image. .

  In so-called metabolic syndrome such as hypertension, dyslipidemia, diabetes and arteriosclerosis, which are typical lifestyle-related diseases, blood pressure, lipid and blood glucose can be easily measured and self-examination is possible. However, at present, arteriosclerosis measurement (commonly known as blood vessel age test) is specialized in circulatory organs equipped with testing equipment such as PWV test (pulse wave velocity), ABI test (ankle brachial blood pressure ratio), and carotid echocardiography. Measurement is not possible if it is not a clinic. On the other hand, a technique for simply measuring hypertension and arteriosclerosis by measuring the fundus blood vessel diameter using a fundus camera owned by an ophthalmology clinic is also known. The fundus is the only place in the human body where blood vessels can be directly observed. If it is assumed that hardening of the fundus artery proceeds in the same manner as the arteries throughout the body, it is possible to measure arteriosclerosis by taking fundus photographs. Fundus cameras are becoming smaller and less expensive, and overseas, special lenses that can be photographed with smartphones are already being sold, and infrastructure for self-photographing of fundus photographs is being prepared.

  Scheie classification has been proposed as a method for measuring hypertension and arteriosclerosis using fundus photographs. The main measurement items are the arteriovenous diameter ratio, the arteriovenous intersection ratio, and the arterial reflex ratio. A measurement support method corresponding to these has also been proposed (see Patent Document 1). In addition, a method of fully automatically measuring the arteriovenous diameter ratio has been proposed (see Patent Document 2).

JP 2007-319403 A Japanese Patent Laid-Open No. 10-243924

  In the technique described in Patent Document 1, regarding the measurement of the blood vessel width, a point where the difference in pixel value from the adjacent pixel is large is regarded as a boundary from the center instructed in a circle by the user, and the distance to the nearest boundary point is calculated. Based on this, the radius of the blood vessel is calculated. However, when the blood vessel boundary is blurred due to the condition of the fundus, there is a problem that the boundary point cannot be specified and the blood vessel width is measured to be small. In the technique described in Patent Document 2, there is a difference in brightness between an artery and a vein (both brightness is lower than the background, but the vein is significantly lower than the artery), and binarization is performed with the same threshold value to extract blood vessels. Then, the artery width is small and the vein width is large. Therefore, the arteriovenous diameter ratio cannot be measured correctly. Also, a method of automatically performing arteriovenous discrimination has been proposed and a function of automatically correcting an erroneous determination location based on the continuity of a blood vessel has been proposed.

Due to the problems as described above, at present, a method in which an ophthalmologist applies a gauge on a fundus photograph and visually measures it is common. Therefore, the arteriovenous diameter ratio is a rough value such as 1/2 or 1/3, and there is a problem that the degree of progress cannot be compared quantitatively. On the other hand, even if the reading accuracy is increased, the variation due to the subjectivity of the measurer and the variation due to the measurement location increase, and it is necessary to sample a large number of measurement locations, but the workload of the ophthalmologist is large and not realistic. In addition, as the blood flow width ratio, there is a demand to obtain not only the arteriovenous diameter ratio between the artery and vein but also the blood column reflection width ratio in the artery. The blood column reflex is a phenomenon in which the central part of the blood vessel reflects more vividly than the blood flow reflex due to the hardening of the blood vessel wall, and the increase in the reflex width directly reflects arteriosclerosis (while, The arteriovenous diameter ratio reflects hypertension based on arteriosclerosis and includes hypertension not caused by arteriosclerosis, so it is not possible to determine arteriosclerosis alone. As arteriosclerosis progresses, the proportion of the blood column reflection width in the blood flow width becomes wider, and the blood column reflection portion changes from bright red to copper / gold when photographed in full color.

  Accordingly, an object of the present invention is to provide a fundus image processing apparatus capable of objectively measuring the blood column reflection width ratio while solving the problem of variation due to the subjectivity of the measurer and the measurement location.

In order to solve the above problems, in the first aspect of the present invention,
An apparatus for processing a fundus image,
Display means for displaying the fundus image and a measurement graphic having a trapezoidal outer shape;
Based on an instruction from the outside, a measurement range setting unit that sets at least one side of two parallel sides of the measurement graphic as a measurement range based on the measurement graphic arranged to overlap the fundus image;
Sampling the sides set as the measurement range, creating a one-dimensional luminance array that is an array of luminance values corresponding to each sample obtained by sampling,
Create a differential array that is an array of differential luminance values using the difference value with the luminance value of the adjacent sample using the luminance array,
Search for the maximum sample taking the maximum value and the minimum sample taking the minimum value for the differential array,
A second minimum sample having a minimum value in the range from the maximum sample to the midpoint is searched for the differential array, and a second maximum sample having a maximum value in the range from the minimum sample to the midpoint is searched. ,
Identifying two end points of the blood flow region based on the maximum sample and the minimum sample, and specifying two boundary points of the blood column reflection region based on the second maximum sample and the second minimum sample Means,
There is provided a fundus image processing apparatus characterized by comprising:
Here, the outer shape of the measurement figure is a trapezoid, but the trapezoid means that at least a pair of opposing sides are parallel and includes a rectangle whose four vertices are perpendicular.

  According to the first aspect of the present invention, based on an instruction from the outside, at least one side of two parallel sides of the measurement graphic is set as a measurement range based on the measurement graphic arranged to overlap the fundus image. Sampling a side set as a measurement range, creating a one-dimensional luminance array that is an array of luminance values corresponding to each sample obtained by sampling, and using the luminance array, Create a differential array that is an array of differential luminance values using the difference values of, search for the maximum sample that takes the maximum value and the minimum sample that takes the minimum value for the differential array, and medium to maximum for the differential array The second minimum sample having the minimum value in the range up to the point is searched, the second maximum sample having the maximum value in the range from the minimum sample to the midpoint is searched, and the maximum sample and the minimum sample are determined. Since the two end points of the blood flow region are specified and the two boundary points of the blood column reflection region are specified based on the second maximum sample and the second minimum sample, the luminance value is small in the blood flow region. The vicinity can be specified as a boundary point, and the blood flow region and the blood column reflection region for measuring the blood flow width can be accurately specified while solving the subjectivity problem of the measurer. In particular, when both of the two parallel sides of the measurement graphic are set as the measurement range, the two end points of the two blood flow regions and the blood column reflection region are based on each of the two parallel sides of the measurement graphic. Since the two boundary points can be specified at the same time, it is possible to reduce variations due to the measurement points by obtaining the average value of the two measurement values. In addition, in order to specify the blood flow region, processing is performed using a one-dimensional array without performing processing in a two-dimensional direction. The end point can be specified.

  In the second aspect of the present invention, the distance between the two boundary points of the two blood column reflection regions specified based on the measurement graphic and the two blood flow regions specified based on the measurement graphic are 2. It further has a blood column reflection width ratio calculating means for calculating a blood column reflection width ratio based on the ratio to the distance between the end points.

  According to the second aspect of the present invention, the distance between the two boundary points of the two blood column reflection regions identified based on the measurement graphic and the two blood flow regions identified based on the measurement graphic are 2. Since the blood column reflection width ratio is calculated based on the ratio to the distance between the end points, the blood column reflection width ratio can be accurately calculated.

  Further, in the third aspect of the present invention, an image for trimming a designated area of the fundus image based on designation from the outside and performing enlargement processing of the fundus image so that the area has a predetermined number of pixels. The image processing apparatus further includes an enlargement unit.

  According to the third aspect of the present invention, based on designation from the outside, the designated region of the fundus image is trimmed, the enlargement process is performed so that the region has a predetermined number of pixels, and the enlarged fundus image Since the measurement range is set for the two blood flow widths of the artery and the intra-arterial reflex, which are calculated in units of the number of pixels, a large integer value is obtained. When calculating the reflection width ratio, it is possible to reduce a rounding error due to division of integer values.

In the fourth aspect of the present invention, the blood flow region end point specifying means includes:
When creating the luminance array, if the fundus image is a color image in which each component of RGB has a predetermined gradation, the gradation value of the G component and the gradation value of the B component are inverted with respect to the fundus image. The grayscale image is converted into a grayscale image by the square root of the product with the value, and a negative / positive-inverted gradation value is given as the luminance value.

  According to the fourth aspect of the present invention, when the luminance array is created, when the fundus image is a color image in which each component of RGB has a predetermined gradation, the gradation value of the G component and B Converted to a grayscale image by the square root of the product with the inverted value of the tone value of the component, and further given the tone value as a negative-positive inverted value as the luminance value, thus improving the contrast to the background of the blood flow region, It is possible to accurately identify the end points of the blood flow region.

  In the fifth aspect of the present invention, the measurement graphic includes a reference line connecting midpoints of two parallel sides of the trapezoid.

  According to the fifth aspect of the present invention, since the measurement graphic is provided with a reference line connecting the midpoints of two parallel sides of the trapezoid, an operation for the user to set the measurement graphic along the blood vessel running In addition, the two parallel sides of the set measurement graphic are orthogonal to the blood vessel traveling direction, and the blood flow width can be measured with high accuracy based on the two end points of the two blood flow regions. Become.

In the sixth aspect of the present invention, the blood flow region end point specifying means includes:
When creating a luminance array,
Sample one of two parallel sides of the measurement graphic, specify two points where each sample is moved by a predetermined distance in two directions orthogonal to the side, and brightness at the two specified points The luminance value corresponding to each sample is corrected using the value.

  According to the sixth aspect of the present invention, when creating the luminance array, one of the two parallel sides of the measurement graphic is sampled, and each sample is moved by a predetermined distance in two directions orthogonal to the side. 2 points are specified, and the luminance value corresponding to the sample is corrected using the luminance values at the two specified points, so that it is spot-like on one of the two parallel sides of the measurement figure. Even if there is an uneven brightness, the influence can be mitigated and correct measurement can be performed.

In the seventh aspect of the present invention,
The blood flow region end point specifying means includes:
When the differential brightness value is smaller than the maximum sample (imax) and larger than the first predetermined value (H _max ), the maximum brightness (i _p ) is closer to the first sample (i = 0) than the first predetermined value. A sample having a small differential luminance value is identified as the first end point (i1) of the blood flow region;
Samples near the end as viewed from the minimum sample (imin) from the differential luminance value is greater the first predetermined value smaller than the second predetermined value (H _min) than the differential luminance value is smaller minimum point _{(i n) (i = s} ) A sample having a differential luminance value greater than a second predetermined value is identified as the second end point (i4) of the blood flow region,
A differential luminance value that is smaller than the third predetermined value and closer to the first sample when viewed from the maximum point (i _p2 ) that has a differential luminance value smaller than the second maximum sample (imax2) and larger than the third predetermined value (H _max2 ). A sample having a value is identified as the first boundary point (i3) of the blood column reflection region;
The second minimum sample (imin2) from the differential luminance value is larger third predetermined value smaller than the fourth predetermined value (H _min2) than the differential luminance value is smaller minimum point (i _n2) fourth at the end of the sample toward viewed from A sample having a differential luminance value larger than a predetermined value is specified as the second boundary point (i2) of the blood column reflection region.

According to the seventh aspect of the present invention, the top (i = 0) when viewed from the maximum point (i _p ) where the differential luminance value is smaller than the maximum sample (imax) and larger than the first predetermined value (H _max ). A sample having a differential luminance value closer to the first sample than the first predetermined value is specified as the first end point of the blood flow region, and a second predetermined value having a differential luminance value larger than the minimum sample (imin) and smaller than the first predetermined value. (H _min) than the differential luminance value is smaller minimum point as viewed from (i _n) the second predetermined value in the sample near the end _{(i = s) (H min} ) is greater than the sample taking the derivative luminance value of the blood flow region It is specified as the second end point, and is closer to the first sample when viewed from the maximum point (i _p2 ) having a differential luminance value smaller than the second maximum sample (imax2) and larger than the third predetermined value (H _max2 ). 3 Take a differential luminance value smaller than a predetermined value The sample is specified as the first boundary point (i3) of the blood column reflection region, and the differential luminance value is greater than the fourth predetermined value (H _min2 ) having a differential luminance value larger than the second minimum sample (imin2) and smaller than the third predetermined value. Since the sample having a differential luminance value larger than the fourth predetermined value near the last sample when viewed from the minimum point (i _n2 ) where is small is specified as the second boundary point (i2) of the blood column reflection region, It is possible to accurately specify the blood flow region and the blood column reflection region of the blood vessel included in the measurement range.

  In the eighth aspect of the present invention, the display means is specified for each of the parallel two sides and a trapezoid whose apex is the four end points of the blood flow region specified for each of the two parallel sides. In addition, a figure obtained by combining trapezoids having apexes at the four boundary points of the blood column reflection region is superimposed on the fundus image and displayed.

  According to the eighth aspect of the present invention, a trapezoid whose apex is the four end points of the blood flow region specified on each of the two parallel sides and the four boundaries of the blood column reflection region specified on each of the two parallel sides A figure composed of a trapezoid with a vertex at a point is displayed overlaid on the fundus image, so it is possible to visually check whether the measurement figure properly indicates the blood flow area and blood column reflection area of the blood vessel. It becomes possible to do.

  According to a ninth aspect of the present invention, there is provided a program for causing a computer to function as the fundus image processing apparatus. The computer includes a computer that performs image processing on a fundus image captured by an image input device built in the same housing.

  According to the ninth aspect of the present invention, the fundus image processing apparatus can be realized by a general-purpose computer by incorporating a program. When a program called an application is installed in a portable computer such as a tablet or a smartphone, image processing can be performed on the fundus image captured by the built-in camera, and the fundus image input device and the fundus image processing device Both functions can be realized by a single computer.

In the tenth aspect of the present invention,
A display device for displaying a measurement graphic when specifying a blood flow region of a blood vessel included in a fundus image,
Based on an instruction from the outside, the external shape is a trapezoid, and a measurement figure with a reference line connecting the midpoints of two parallel sides is displayed.
Two parallel sides of the graphic for measurement arranged based on an instruction from the outside are set as measurement ranges, and two boundaries of the blood flow region and two boundaries of the blood column reflection region are set based on the two sides of the set measurement range. A figure obtained by combining a trapezoid having four apexes with four end points of the specified blood flow region and a trapezoid having four apex points with the four boundary points of the specified blood column reflection region is identified. A graphic display device for measurement, which is characterized by displaying instead of the above, is provided.

  According to the tenth aspect of the present invention, when the blood flow region of the blood vessel included in the fundus image is specified, the reference line connecting the midpoints of the two parallel sides having an outer shape based on an instruction from the outside A measurement graphic having a measurement area is displayed, two parallel sides of the measurement graphic arranged based on an instruction from the outside are set as measurement ranges, and 2 of the blood flow region is set based on the two sides of the set measurement ranges. Identify two boundary points between the end point and the blood column reflection region, and combine a trapezoid with the four end points of the specified blood flow region as the four vertices and a trapezoid with the four boundary points in the specified blood column reflection region as the vertices Since the figure to be displayed is displayed in place of the figure for measurement, the measurement figure is deformed like a so-called measurement gauge and gives an impression that the blood flow area has been measured. Intuitively determine blood flow width and blood column reflection width in the blood column reflection region Grip Shi easily.

  According to the present invention, it is possible to objectively measure the blood column reflection width ratio while solving the problem of variation due to the subjectivity of the measurer and the measurement location.

1 is a hardware configuration diagram of a fundus image processing apparatus 100 according to an embodiment of the present invention. 1 is a functional block diagram illustrating a configuration of a fundus image processing apparatus according to an embodiment of the present invention. It is a flowchart which shows the process outline | summary of the fundus image processing apparatus according to an embodiment of the present invention. It is a figure which shows a fundus image and an analysis range. It is a figure which shows the initial state of the figure for a measurement. It is a figure which shows the mode of a deformation | transformation by operation of the figure for a measurement. It is a figure which shows an analysis range, the figure for a measurement, and a measurement range setting screen. It is a flowchart which shows the specific process of the blood-flow area | region end point and blood column reflection area | region boundary point in step S400. It is a figure which shows the figure based on the figure for a measurement, and a measurement result. It is a figure which shows the relationship of the pixel moved by 1.5 pixels and start point edge | side SH or end point edge | side EH. It is a figure which shows the relationship between the fundus image of a measurement range, a luminance value, and a differential luminance value. It is a figure which shows the relationship between the luminance value in case a part of blood vessel of both sides is contained in a measurement range, and a differential luminance value. It is a figure which shows the example which displayed the figure after measurement on the actual image. It is a figure which shows the example by which the figure for a measurement is arrange | positioned in three arteries within the analysis range. It is a figure which shows the example by which the figure for a measurement is arrange | positioned in three arteries within the analysis range.

DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the drawings.
<1. Device configuration>
FIG. 1 is a hardware configuration diagram of a fundus image processing apparatus according to an embodiment of the present invention. The fundus image processing apparatus 100 according to the present embodiment can be realized by a portable general-purpose computer having a camera and a wireless communication function such as a tablet or a smartphone. As shown in FIG. 1, the fundus image processing apparatus 100 according to the present embodiment includes a CPU (Central Processing Unit) 1, a RAM (Random Access Memory) 2 that is a main memory of a computer, and programs and data executed by the CPU 1. A large-capacity storage device 3 such as a flash memory for storing data, a touch panel input I / F (interface) 4 for moving / deforming a measurement figure, which will be described later, and instructing an operation menu, and wireless A data input / output I / F (interface) 5 for inputting images from an external fundus camera via communication or the like, and outputting analysis results to an external server computer or printer, and a liquid crystal display for displaying image data, etc. A display unit 6 that is a display device such as an organic EL display and a built-in camera (for photographing the fundus) And an image input unit 7 that can be used as a fundus camera by mounting a light source and a macro lens, and are connected to each other via a bus. Of course, any computer having a similar function can be realized by a stationary general-purpose computer.

  FIG. 2 is a functional block diagram illustrating a configuration of the fundus image processing apparatus according to the present embodiment. In FIG. 2, 10 is an instruction input means, 20 is a display means, 30 is an image enlargement means, 40 is a measurement range setting means, 50 is a blood flow region end point specifying means, 60 is a blood column reflection width ratio calculation means, and 70 is a fundus oculi. Image storage means 80 is an information storage means.

  The instruction input means 10 transmits an instruction input from the outside to each means, and is realized by a touch panel input I / F (interface) 4 and the CPU 1. The display unit 20 is for displaying information stored in the storage unit and information processed by each unit, and is realized by the display unit 6 and the CPU 1. The image enlarging means 30 trims the set analysis range in the fundus image read from the fundus image storage means 70, and performs enlarging processing so as to obtain a predetermined number of pixels. This is because the blood flow width of the artery to be measured is narrower than that of the vein, and it is necessary to measure a minute blood column reflection width within the blood flow width, and these measurements are performed in pixel units. However, the value is too small if the number of pixels in the original image is kept, and dividing the blood column reflection width by the blood flow width does not provide sufficient accuracy to determine the degree of hypertension or arteriosclerosis due to rounding error. It is. The measurement range setting means 40 sets at least one side of two parallel sides of the measurement graphic as the measurement range based on the measurement graphic arranged in the range including the artery of the fundus image. The blood flow region end point specifying means 50 specifies the two end points of the blood flow region, which is a region corresponding to the blood flow, from the measurement range. Note that a maximum of two measurement ranges, which are one-dimensional blood flow regions, are specified simultaneously based on two parallel sides from a single measurement graphic.

  The blood column reflection width ratio calculating means 60 calculates the distance between the two boundary points of the two blood column reflection areas specified based on the measurement graphic and the two blood flow areas specified based on the measurement graphic. The blood column reflection width ratio is calculated based on the ratio with the distance between the end points. Actually, from the measurement range corresponding to the artery, each of two parallel sides of the measurement graphic corresponds to the measurement range, and two blood flow regions and blood column reflection regions are specified, and the two distances are determined. Since it is calculated, the blood column reflection width ratio is calculated based on the ratio between the total values of the two distances. Here, in order to calculate the two end points of the one-dimensional blood flow region and the two boundary points of the blood column reflection region, the reason for setting the one-dimensional measurement range using the two-dimensional measurement graphic is as follows. When the measurement range is set by specifying on the screen with a one-dimensional straight line, the blood flow region becomes oblique to the traveling direction of the blood vessel to be measured. Because it tends to be.

  The “blood column reflection width ratio” is a kind of blood flow width ratio, and is a ratio of a region where a part of the blood flow width has different optical reflection characteristics. There is no physically different tissue called “blood column” in the artery or blood flow, and the blood vessel wall becomes thicker due to arteriosclerosis, so that the reflection characteristic of the blood flow width changes partially brightly. The boundary of the blood column reflection width is not as clear as the boundary of the blood flow width. As arteriosclerosis progresses, the entire blood flow width shows blood column reflection, and the reflected light from the white light source of the fundus camera becomes a vivid copper or gold color. In the present invention, the “blood flow” region is specified and the “blood flow” width ratio is calculated because the blood vessel width is transparent, which makes it difficult to specify the blood vessel wall and calculate the blood vessel width ratio. Because. However, arteriosclerosis generally makes the blood vessel wall thicker and the blood flow width narrower, but the blood vessel width does not change much. Therefore, rather than calculating the blood vessel width ratio, the method of specifying the blood flow region and calculating the blood flow width ratio is more useful for determining the degree of hypertension or arteriosclerosis. However, in some cases, the terms “blood vessel” and “blood vessel width” are easier to understand intuitively. Therefore, the terms “blood vessel” and “blood vessel width” are also used below.

  The image enlarging means 30, the measurement range setting means 40, the blood flow region end point specifying means 50, and the blood column reflection width ratio calculating means 60 are realized by the CPU 1 executing a program stored in the storage device 3. The fundus image storage means 70 shoots in full color (in the case of visible light) or monochrome (in the case of infrared light) using a visible light or infrared / light source type fundus camera by the data input / output I / F 5 or the image input unit 7. The storage unit 3 stores the full-color fundus image to be extracted from the blood flow region. Note that a fundus image captured by an infrared light source type fundus camera is normally captured in monochrome, but is stored in the storage device 3 as a full-color image having the same value of RGB. In addition, the infrared light source type fundus camera is externally attached to the computer and is taken into the storage device 3 via the data input I / F 5. However, the visible light source type fundus camera is incorporated in the computer. It is also possible to use the fundus camera by attaching a light source and a macro lens for fundus photography to the image input unit 7. When the fundus image processing apparatus reads a full-color fundus image and performs the processing as it is, the RAM 2 serves as the fundus image storage unit 70. A full-color fundus image is image data recorded with three RGB components, and is obtained by photographing the fundus of a subject. In the present embodiment, RGB recorded with 8-bit 256 gradations for each color is used. The information storage unit 80 stores information on the blood flow region end point specified by the blood flow region end point specifying unit 50, the blood flow width calculated by the blood column reflection width ratio calculating unit 60, the blood flow width ratio, and the like. And is realized by the storage device 3.

  Each component shown in FIG. 2 is actually realized by installing a dedicated program in hardware such as a computer and its peripheral devices as shown in FIG. That is, the computer executes the contents of each means according to a dedicated program. If the camera built in the computer is equipped with a light source or macro lens for fundus photography, the function of the fundus camera can be replaced (currently not approved in Japan), and it is consistently single from fundus photography to fundus image processing. It can be realized with a portable computer. Note that in this specification, a computer means an apparatus having an arithmetic processing unit such as a CPU and capable of data processing, and is not limited to a portable terminal such as a tablet or smartphone having a camera or a wireless communication function. The concept also includes general-purpose computers such as personal computers.

  A dedicated program for operating the CPU 1 and causing the computer to function as a fundus image processing apparatus is installed in the storage device 3 illustrated in FIG. By executing this dedicated program, the CPU 1 functions as an image enlarging means 30, a measurement range setting means 40, a blood flow region end point specifying means 50, a blood column reflection width ratio calculating means 60, and an instruction input means 10, A function as a part of the display means 20 is realized. The storage device 3 not only functions as the fundus image storage unit 70 and the information storage unit 80 but also stores various data necessary for processing as the fundus image processing device.

<2. Processing action>
<2.1. Pretreatment>
First, a fundus image to be processed is taken or prepared. Ideally, the fundus image for the purpose of blood vessel measurement is a gray scale (monochrome) image (fundus angiographic image) captured by an infrared light source. However, since the photographing apparatus is expensive, in this embodiment, a color image photographed with a popular visible light source is set as a processing target. A gray scale (monochrome) image is also unified into a full color image format. A color image photographed with a visible light source is also converted into a gray scale image later, as will be described later. This is because a blood flow region can be extracted with high accuracy by utilizing color features. The color image is preferably a full color image having at least 8 bits and 256 gradations or more for each of the R, G, and B components. As a full-color fundus image, if there is an image file taken in full color by a digital fundus camera, it can be used as it is. For past records recorded on photographic media using an analog fundus camera, the stored analog monochrome or color negative / positive film, photographic paper, instant photo, etc. are read in full color with a scanner, etc. Get the fundus image file. At this time, even if the original image is monochrome, it is read in full color. Recently, a fundus image is obtained in the form of a digital file by photographing in full color using a digital fundus camera of a visible light source. The acquired fundus image is stored in the fundus image storage unit 70 of the fundus image processing apparatus. In the present embodiment, a full-color image having 8 bits and 256 gradations for each of the R, G, and B components is prepared as a fundus image.

<2.2. Process Overview>
Next, the processing operation of the fundus image processing apparatus shown in FIGS. 1 and 2 will be described. FIG. 3 is a flowchart showing an outline of processing of the fundus image processing apparatus according to the embodiment of the present invention. First, the image enlarging means 30 sets an analysis range (step S100). Specifically, first, the fundus image is displayed on the display unit 6 to prompt the user to specify the analysis range. When the user designates a predetermined range in the fundus image via the touch panel input I / F 4 that is the instruction input means 10, the image enlargement means 30 sets the designated range as the analysis range. The predetermined range can be specified by, for example, a user performing a drag operation or the like on the rectangle displayed on the fundus image using the touch panel or the like to specify the analysis range. A display example of the fundus image at this time is shown in FIG. The example of FIG. 4 shows a state in which an analysis range of 46 pixels wide by 49 pixels high is set in a fundus image of 1280 pixels wide × 1024 pixels high. The analysis range is set as a range including all of a plurality of measurement ranges set later. That is, the measurement range is set within the set analysis range.

  When the analysis range is set, the image enlarging means 30 trims the set analysis range and performs an enlargement process on the trimmed portion (step S200). Specifically, the number of pixels of the fundus image in the analysis range is increased. At this time, although the resolution is apparently increased, it is not necessary to add special interpolation to the enlargement process to improve the resolution itself. The number of pixels of the image to be processed is preferably set so that the blood column reflection width within the blood flow width of the artery is about several tens of pixels. When the entire fundus occupies an image of horizontal 1280 pixels × vertical 1024 pixels, the blood column reflection width within the blood flow width may be less than one pixel, and analysis cannot be performed as it is. Therefore, when the original fundus image is 1280 horizontal pixels × 1024 vertical pixels, the image is enlarged by 600% to 900% in both vertical and horizontal directions. Here, the enlargement ratio is set to 900% in advance. In the present embodiment, since it is not necessary to increase the resolution, an appropriate pixel value may be given to the increased pixels in order to perform the enlargement process, and it is not necessary to use an advanced interpolation method. Therefore, in this embodiment, a known bilinear method is used.

  When the analysis range is trimmed and enlarged, the measurement range is set (step S300). At this time, first, the measurement range setting means 40 prompts the user to specify the measurement range. Specifically, display is performed in which measurement figures having an outer shape of a trapezoid (meaning including rectangles and squares) are arranged on the enlarged analysis range image. The measurement graphic is a graphic that gives an impression like a measurement gauge when viewed from the user. When the measurement graphic is displayed, the user scales, moves, and rotates the measurement graphic vertically and horizontally by the instruction input unit 10 and arranges the parallel two sides across the blood vessel width to be measured.

  FIG. 5 shows a measurement graphic displayed on the display unit 6. As shown in FIG. 5, the measurement graphic is a rectangle composed of four vertices of a vertex 11, a vertex 12, a vertex 21, and a vertex 22. The midpoint of the line segment connecting the vertices 11 and 12 is called the start point, the midpoint of the line segment connecting the vertices 21 and 22 is called the end point, and the line segment connecting the start point and the midpoint is the reference line KH. Further, a line segment having the start point as the midpoint is set as the start point side SH, and a line segment having the end point as the midpoint is set as the end point side EH.

  When one of the six points on the measurement graphic is selected by the instruction input means 10, the point becomes movable, and is fixed when an operation instruction is given again at the designated position. When the top of the measurement graphic is selected, the entire measurement graphic is movable. When the start point or end point of the measurement figure is selected, the length can be expanded and contracted and rotated. In the example of FIG. 6A, a case where an end point surrounded by a circle is selected is shown. When the end point is selected, as shown in the upper diagram of FIG. 6A, the length of the entire measurement figure is expanded or contracted in the right direction by an operation instruction in the right direction. When the end point is selected, as shown in the lower diagram of FIG. 6A, the entire measurement figure is rotated around the start point by an operation instruction in the vertical direction. In FIG. 6A, the points marked with x indicate points that are fixed and do not move. FIG. 6A shows the case where the end point is selected, but it is possible to expand / contract / rotate similarly when the start point is selected. In the example of FIG. 6A, the case where the end point is selected has been described. However, even when the start point is selected, the length can be expanded and contracted in the same manner.

  When the vertex 11 or the vertex 12 of the measurement figure is selected, that is, when either one of both ends of the start point side SH is selected, the width is set by an operation instruction in the vertical direction as shown in FIG. Perform bulk expansion and contraction. When the vertex 21 or the vertex 22 of the measurement figure is selected, that is, when either one of both ends of the end point side EH is selected, as shown in FIG. Scale the width of the rectangle only. When the vertex 21 or the vertex 22 of the measurement graphic is selected, a mode is such that another narrow rectangle appears inside the measurement graphic that has been a single rectangle. A narrow rectangle has a pair of two parallel sides that overlap the start and end sides of the outer rectangle, so it appears that only two sides parallel to the reference side KH have appeared. Note that the width of only the end point side EH is expanded or contracted. In FIGS. 6B and 6C, the points marked with x indicate points that are fixed and do not move. This operation is a function provided so that the measurement graphic can be manually measured according to the blood vessel width, and is often not used because the subjectivity of the operator is included in the measurement result.

  In the initial state as shown in FIG. 5, first, the user moves the entire measurement graphic to the position of the blood vessel to be measured. Then, the entire measurement graphic is rotated so that the reference line KH is approximately aligned with the center line of the blood vessel traveling to be measured (virtual curve connecting the center of the blood vessel width in the traveling direction). At this time, the length of the entire measurement graphic is also expanded and contracted so that the center line included in the measurement graphic is substantially a straight line and substantially coincides with the reference line KH. Finally, the width is expanded and contracted so that the side connecting the vertex 11 and the vertex 21 and the side connecting the vertex 12 and the vertex 22 protrude sufficiently outside the blood vessel to be measured. At this time, it is desirable that only the blood flow region to be measured is included in the measurement graphic. However, even if some blood flow regions to be measured are included in the measurement graphic, the measurement is hindered. Absent. Then, the start point side SH and the end point side EH are arranged at a position where the blood vessel intersects the blood vessel substantially at a right angle. The blood flow is usually not a straight line but a curved line. For this reason, “substantially a right angle” is not meant to be strictly 90 degrees, but includes an angle that allows the blood flow width to be appropriately measured.

  When the measurement graphic is arranged at a predetermined position of the analysis range in the fundus image, the measurement range setting means 40 sets each of two parallel sides of the measurement graphic as a one-dimensional measurement range based on the designated measurement graphic. Set. A display example of the analysis range and the measurement graphic at this time is shown in FIG. In the example of FIG. 7A, a measurement figure having a rectangular outer shape is arranged on a fundus image of 414 pixels wide × 441 pixels vertical, which is an analysis range enlarged by 900%. As will be described later, the measurement is performed using pixels on two sides of the start point side SH and the end point side EH and in the vicinity thereof. Therefore, by arranging a two-dimensional measurement graphic, the start point side SH and the end point side EH of the measurement graphic are set as a one-dimensional measurement range. If a one-dimensional line segment is specified on the screen to set a one-dimensional measurement range, the measurement range orthogonal to the blood vessel cannot be set appropriately. On the other hand, the measurement range orthogonal to the blood vessel can be appropriately set by arranging and designating a two-dimensional measurement graphic having opposite sides parallel to each other on the screen. In this embodiment, two parallel sides of one measurement figure are set as measurement ranges. However, only one of the two parallel sides can be set as the measurement range. In the case of measuring the blood column reflection width ratio, measurement is possible even with only one side, but since the width and inclination of the blood vessel are not always constant, two parallel sides are measured. It is possible to perform more accurate measurement.

  When the measurement range is set, the measurement range setting means 40 displays a measurement range setting screen and prompts the user to select whether the measurement range is an artery or a vein. FIG. 7B shows an example of the measurement range setting screen. The measurement range setting means 40 instructs the user to specify a measurement range based on a plurality of measurement figures and to select either an artery, a vein, or an artery / blood column reflection for each of the measurement figures arranged. Prompt and set the analysis conditions for the measurement range. Therefore, the same setting is performed for two measurement ranges corresponding to one measurement graphic. When determining the blood column reflection width ratio, the measurement range is set so that at least one is set at a location corresponding to the artery.

  In the measurement range setting screen shown in FIG. 7B, not only a selection instruction as to whether or not the measurement range is an artery / blood column reflex but also various analysis conditions for the measurement range can be specified. However, it is indispensable to specify whether or not it is an artery / blood column reflex.

  Once the measurement range is set, the blood flow region end point specifying means 50 next specifies the blood flow region end point and the boundary point of the blood column reflection region according to the contents input using the measurement range setting screen. This is performed (step S400). Specifically, the end point of the blood flow region and the boundary point of the blood column reflection region are specified using the start point side SH and the end point side EH, which are measurement ranges.

  FIG. 8 is a flowchart showing the specific processing of the end point of the blood flow region in step S400. First, the blood flow region end point specifying means 50 acquires information on the measurement figure corresponding to the two sides of the start point side SH and the end point side EH that are the measurement range (step S410). Specifically, first, the coordinates of the starting point (Xo1, Yo1) and the coordinates of the end point (Xo2, Yo2), the vertices (Xo11, Yo11) and (Xo12, Yo12) of the starting point side SH, and the end point side EH are measured. The vertices (Xo21, Yo21) and (Xo22, Yo22) are acquired, and ½ of the larger length of the start point side SH or the end point side EH is acquired as the width W of the graphic for measurement. Start point coordinates (Xo1, Yo1) and end point coordinates (Xo2, Yo2), start point side SH vertices (Xo11, Yo11) and (Xo12, Yo12), end point side EH vertices (Xo21, Yo21) and (Xo22, Yo22) ) Specifies the pixel position of the fundus image. FIG. 9A is a diagram illustrating the measurement graphic arranged on the blood vessel of the fundus image and the coordinate values of each point. As shown in FIG. 9A, the measurement graphic has a shape having a line segment connecting the midpoint of the start point side SH and the midpoint of the end point side EH inside the rectangle. A line segment inside the rectangle is a reference line KH, which is a guideline for the user to place a measurement graphic on the fundus image, and is not used for actual measurement processing.

  As described above, the line segment obtained by connecting the two vertices (Xo11, Yo11) and (Xo12, Yo12) on the start point side connects the start point side SH and the two vertices (Xo21, Yo21), (Xo22, Yo22) on the end point side. The line segment obtained in is the end point side EH. Then, the following steps S420 to S470 are executed for each of the start point side SH and the end point side EH.

  First, a luminance array is created (step S420). Specifically, first, a side (a start point side SH or an end point side EH) arranged to intersect with a blood vessel is sampled. That is, the coordinates of one vertex of the side (starting point side SH or end point side EH) arranged intersecting the blood vessel is (Xs, Ys), the coordinate of the other vertex of the side is (Xe, Ye), and each side S + 1 samples (X (i), Y (i)) of i = 0,..., S are set at predetermined intervals. Specifically, when the processing target side is the start point side SH, Xs = Xo11, Ys = Yo11, Xe = Xo12, Ye = Yo12, and when the processing target side is the end point side EH, Xs = Xo21, Ys = Yo21, Xe = Xo22, Ye = Yo22. The coordinates (X (i), Y (i)) of each sample are Dx = Xe−Xs, Dy = Ye−Ys, and X (i) = Dx · i / s + Xs, Y (i) = Dy · i / It is defined by s + Ys. Thus, a one-dimensional coordinate array (X (i), Y (i)) (i = 0,..., S) is created. Subsequently, the luminance value of the pixel corresponding to each sample is acquired, and a one-dimensional luminance array G (i) (i = 0,..., S) composed of s + 1 luminance values is obtained. At this time, in order to further improve accuracy, a point moved by 1.5 pixels from each sample is set on both sides in the orthogonal direction with respect to the side. The reason for 1.5 pixels is that even if the sides are inclined at a maximum of 45 °, the luminance values of pixels different by one or more pixels are acquired. The length to be moved is not limited to 1.5 pixels, and can be set as appropriate as long as it is 1.5 pixels or more. FIG. 10 is a diagram illustrating the relationship between the start point side SH or the end point side EH and the pixel moved by 1.5 pixels.

  In FIG. 10, the central line segment is the start point side SH or the end point side EH, and s + 1 samples (X (i), Y (i)) are set from (Xs, Ys) to (Xe, Ye). ing. A shaded portion surrounded by a broken line indicates a blood flow region. Then, for each sample (X (i), Y (i)), the points (Xl, Yl), (Xr, Yr) are moved to the positions moved by 1.5 pixels in the direction orthogonal to the start point side SH or the end point side EH. ) Is set. For each sample (X (i), Y (i)), two points moved by 1.5 pixels from the start point side SH or end point side EH by setting the points (Xl, Yl), (Xr, Yr) The luminance value of the point on the line segment is reflected on the luminance value of the sample of the start point side SH or the end point side EH. Thus, if only the start point side SH or the end point side EH reflects the luminance value of the sample moved by 1.5 pixels from the start point side SH or the end point side EH, the start point side SH or the end point side EH is reflected. This is because if the sample has spot-like luminance unevenness, correct measurement cannot be performed due to the influence. Such points (Xl, Yl) and (Xr, Yr) are calculated by executing processing according to the following [Equation 1].

[Formula 1]
Xl = −Dy · 1.5 + X (i)
Yl = Dx · 1.5 + Y (i)
Xr = Dy · 1.5 + X (i)
Yr = −Dx · 1.5 + Y (i)

  Then, the luminance value of each sample (X (i), Y (i)) is obtained by averaging the luminance values of (X (i), Y (i)) (Xl, Yl), (Xr, Yr). And That is, the luminance value corresponding to each sample (X (i), Y (i)) is corrected using the luminance value at the two specified points (Xl, Yl) and (Xr, Yr). When the fundus image is a full-color image, it is necessary to convert to a gray scale in order to obtain a luminance value. In the present embodiment, gray scale conversion is performed on the target pixel as follows.

  A full-color fundus image is 8-bit 256-gradation image data for each color. Therefore, the fundus image having the number of pixels in the x direction Xg and the number of pixels in the y direction Yg is assumed to be Image (x, y, c) = 0 to 255 (x = 0,..., Xg-1; y = 0,..., Yg-1; c = 0, 1, 2).

  When converting to a grayscale image, only the Green component is used for the full-color fundus image Image (x, y, c), ie, Gr (x, y) = Image (x, y, 1) It is known that the method of conversion is the best. The reason is that the resolution of the Red and Blue components is lower than that of the Green component in the three primary color filters used in the fundus camera, and the outline of the blood vessel tends to be unclear in these color separation images. However, as a result of analysis of many fundus camera images, the present inventor found that the Green component was clear near the center where the illumination light hits sufficiently, but the Blue component rather than the Green component in the periphery where the illumination light was weak. In contrast to the Green component, the inversion phenomenon was found in which the brightness of the blood vessel part was higher than that of the surroundings (the brightness of the blood vessel part was lower than that of the surroundings in the entire fundus image). Therefore, in the present embodiment, a method of performing gray scale conversion by executing processing according to the following [Equation 2] is proposed. However, [Equation 2] is not applied to all full-color fundus images, and the conventional method of converting by Gr (x, y) = Image (x, y, 1) is more appropriate. Since there are not a few fundus images, it is necessary to select these appropriately for each given fundus image. The blood flow region end point specifying means 50 performs grayscale conversion by executing processing according to the following [Equation 2].

[Formula 2]
Gr (x, y) = [Image (x, y, 1). (255-Image (x, y, 2))] ^1/2

  In the above [Equation 2], Image (x, y, 1) represents a G (green, green) component of the full-color fundus image, and Image (x, y, 2) represents the full-color fundus image. Among these, the component of B (blue, blue) is shown. Therefore, in the above [Equation 2], a gray scale image Gr (x, y) is obtained by obtaining a geometric mean value (square root of the product) of the negative and positive inversion of the G component and B component of each pixel. .

  Further, the blood flow region end point specifying unit 50 performs a process according to the following [Equation 3] on the grayscale image Gr (x, y), thereby performing negative / positive reversal to obtain a grayscale fundus image. Gray (x, y) is obtained.

[Formula 3]
Gray (x, y) = 255-Gr (x, y)

  By executing the processing according to the above [Equation 3] and performing negative / positive inversion, the pixel value of the blood flow region having a lower luminance than the surroundings is converted to a high state. Here, the negative / positive inversion means a process of converting a smaller value as the original pixel value (gradation value) is larger.

  In the present embodiment, only the pixels corresponding to (Xl, Yl) and (Xr, Yr) may be grayscaled. When luminance values of (X (i), Y (i)) (Xl, Yl), (Xr, Yr) are obtained by gray scale conversion as described above, processing according to the following [Equation 4] To calculate the luminance array G (i).

[Formula 4]
G (i) = [Gray (X (i), Y (i)) + Gray (X1, Yl) + Gray (Xr, Yr)] / 3

  As described above, each sample has s + 1 set between both ends of the start point side SH or the end point side EH for convenience of calculation. Further, although the start point side SH or the end point side EH may be set to be inclined with respect to the xy axis that is the pixel array, the sample coordinate value is approximated to an integer value, and thus the luminance value of the corresponding pixel as it is. Can be obtained.

  When the luminance value is obtained for each sample i and the luminance array G (i) is obtained, the differential array D (i) converted into the differential luminance value array using the luminance array G (i) is then calculated. (Step S430). Specifically, the differential array D (i) is calculated by executing processing according to the following [Equation 5]. However, when i <3 or i> s−2, D (i) = 0.

[Formula 5]
D (i) = G (i) + G (i + 1) + G (i + 2) −G (i−1) −G (i−2) −G (i−3)

  In the above [Formula 5], samples i + 1 and i-1 adjacent to the sample i are included, and samples i, i + 1, i + 2, i-1, i-2, i-3 are used as a total of six samples. Although the differential luminance value D (i) at i is calculated, the differential luminance value D (i) may be calculated using more samples. In the above [Equation 5], the addition value + G (i + 1) with the adjacent sample and the difference value −G (i−1) are used.

Once the differential array D (i) is obtained, the blood flow region end point specifying means 50 next specifies the maximum value and the minimum value of the differential array D (i) (step S440). Specifically, the maximum value D _max and the minimum value D _min of the differential array D (i) are calculated by executing processing according to the following [Equation 6].

[Formula 6]
D _max = MAX _{i = 0, S / 2} D (i) = D (i _max )
D _min = MIN _{i = S / 2, S} D (i) = D (i _min )

In the above [Expression 6], “MAX _{i = 0, S / 2} ” in the first expression represents the maximum value of D (i) in each sample i from the subscript i = 0 to i = s / 2. Also, in the above [Formula 6], “MIN _{i = S / 2, S} ” in the second formula represents the subscript i = s / 2 to i = s in each sample i. The minimum value of D (i) ”is obtained. The sample imax taking the maximum value _Dmax is specified as the maximum sample, and the sample imin taking the minimum value _Dmin is specified as the minimum sample.

Here, the relationship between the fundus image in the measurement range, the luminance value, and the differential luminance value is shown in FIG. In FIG. 11, the horizontal axis represents the sample i. The upper part of FIG. 11 is a fundus image of the measurement range, and the shaded portion indicates a portion other than the blood column reflection region in the blood flow region. The central part between the shaded parts is the blood column reflection region. That is, the example of FIG. 11 shows a case where the measurement range is set to a range including one blood vessel. The middle part of FIG. 11 shows the luminance value G (i), which indicates that the luminance value is higher toward the center of the blood flow region excluding the blood column reflection region. The lower part of FIG. 11 is a differential luminance value D (i) that is a differential value of the luminance value G (i). In step S440, the maximum value _Dmax and the minimum value _Dmin of the differential luminance value are calculated. In order to measure the blood column reflection width ratio, it is necessary to specify the boundary of the blood column reflection region in the blood flow region. For this purpose, it is necessary to specify a region where the luminance value is small in the blood flow region. The two boundary positions of the region where the luminance value is small are near the position where the differential luminance value takes the second maximum value which is the second largest after the maximum value _Dmax , and the second minimum value where the differential luminance value is the second smallest value after the minimum value _Dmin it is conceivable that. That is, in the order of i = 0 to i = s, the vicinity of the position where the differential luminance value has the maximum value _Dmax is the start position of the blood flow region, and the vicinity of the position where the differential luminance value has the second minimum value is the blood column. The start position of the reflection area, the vicinity of the position where the differential luminance value takes the second maximum value is the end position of the blood column reflection area, and finally the vicinity of the position where the differential luminance value takes the minimum value D _min is the blood flow area. End position.

Once the maximum value D _max and the minimum value D _min of the differential array D (i) are obtained in step S440, the sample median value s / is then determined from the maximum value D _max and the minimum value D _{min of} the differential array D (i). A second maximum value D _max2 and a second minimum value D _min2 closer to 2 are specified (step S450). Specifically, the second maximum value D _max2 and the second minimum value D _min2 are searched by executing processing according to the following [Equation 7].

[Formula 7]
D _min2 = MIN _{i = imax + 1, S / 2} D (i) = D (i _min2 )
D _max2 = MAX _{i = S / 2, imin-1} D (i) = D (i _max2 )

In the above [Expression 7], “MIN _{i = imax + 1, S / 2} ” in the first expression is the second of D (i) in each sample i from the subscript i = imax + 1 to i = s / 2. In other words, the minimum value in the range from the maximum sample imax to the midpoint s / 2 is determined as the second minimum value. Of course, the minimum value is not taken in the maximum sample imax. Therefore, the search itself starts from the next sample imax + 1, and in the above [Expression 7], “MAX _{i = S / 2, imin−1} ” in the second expression is the subscript. It shows that “the second maximum value of D (i) in each sample i from i = s / 2 to i = imin−1” is obtained. That is, the maximum value in the range from the midpoint s / 2 to the minimum sample imin is obtained as the second maximum value. In addition, since it is natural that the maximum value is not taken in the minimum sample imin, the search itself has been completed up to imin−1 which is the adjacent sample. Sample imax2 taking the second maximum value D _max2 second maximum sample, the sample imin2 taking the second minimum value D _min2 is specified as the second minimum sample.

When the maximum value D _max and the minimum value D _min of the differential array D (i) are obtained in step S440, the blood flow region end point specifying means 50 specifies the two end points of the blood flow region (step S460). The process of step S460 can be performed in parallel with the process of step S450 at a stage where the process of step S450 is not completed. Specifically, the two end points of the blood flow region are specified by executing processing according to the following [Equation 8].

[Formula 8]
H _max = D _max · α
_Maximums satisfying D (i _p ) ≧ D (i _p +1) and D (i _p ) ≧ D (i _p −1) and D (i _p ) ≧ H _max in the order from i = s / 2 to i = 0. searching the value _{D (i p), i1 <} i p with D (i1) <identify samples i1 satisfying H _max as a first end point of the blood flow region H _min = D _min · α
_Minimums satisfying D (i _n ) ≦ D (i _n +1), D (i _n ) ≦ D (i _n −1) and D (i _n ) ≦ H _min in the order from i = s / 2 to i = s. searching the value D (i _n), identifying the sample i4 satisfying D (i4)> H _min at i4> i _n a second end point of the blood flow region

In the above [Expression 8], searching for local maximum values in the order of i = s / 2 to i = 0 and searching for local minimum values in the order of i = s / 2 to i = s starts from the sample near the midpoint. Alternatively, the search is performed on the last sample side to indicate the two end points of the blood flow region. In the above [Equation 8], α is a real value from 0 to 1. In the present embodiment, α is set to 0.8. By using the coefficient α, the first predetermined value H _max and the second predetermined value H _min are determined. In the lower part of FIG. 11, the magnitudes of the arrows indicated by D (i1) and D (i4) are differential luminance values, and samples i1 and i4 taking these differential luminance values D (i1) and D (i4) are blood. Two end points of the flow region.

After all, in step S460, the blood flow region endpoint identification unit 50, the sample side of the head (i = 0) as seen from the maximum point is large differential luminance value than the first predetermined value H _max small differential luminance value than the maximum sample The sample having the differential luminance value smaller than the first predetermined value H _max is specified as the first end point of the blood flow region, and the second predetermined value H _{min having a} differential luminance value larger than the minimum sample and smaller than the first predetermined value H _max . A process is performed in which a sample having a differential luminance value larger than the second predetermined value _Hmin near the sample at the end (i = s) as viewed from the minimum point having a smaller differential luminance value is specified as the second end point of the blood flow region. ing.

When the second maximum value D _max2 and the second minimum value D _min2 of the differential array D (i) are obtained in step S450, and the two end points of the blood flow region are specified in step S460, then the blood flow region end point specifying means 50 identifies two boundary points of the blood column reflection region (step S470). Specifically, two boundary points of the blood column reflection region are specified by executing processing according to the following [Equation 9].

[Formula 9]
H _min2 = D _min2 · α
_Minimums satisfying D (i _n2 ) ≦ D (i _n2 +1), D (in ₂ ) ≦ D (in ₂ −1) and D (in ₂ ) ≦ H _min2 in the order of i = i1 + 1 to i = s / 2. The value D (i _n2 ) is searched, and the sample i2 that satisfies D (i2)> H _min2 when i2> i _n2 is specified as the first boundary point of the blood column reflection region H _max2 = D _max2 · α
In order from i = i4-1 to i = s / 2, D (i _p2 ) ≧ D (i _p2 +1) and D (i _p2 ) ≧ D (i _p2 −1) and D (i _p2 ) ≧ H _max2 The maximum value D (i _p2 ) to be satisfied is searched, and the sample i3 that satisfies D (i3) <H _max2 with i3 <i _p2 is specified as the second boundary point of the blood column reflection region

In the above [Equation 9], searching for local minimum values in the order of i = i1 + 1 to i = s / 2, and searching for local maximum values in the order of i = i4-1 to i = s / 2 means that the specified blood flow This indicates that the search is performed from the first end point i1 of the region and the second end point i4 of the blood flow region to the midpoint side of the sample row, and two boundary points of the blood column reflection region are specified. Also, in the above [Equation 9], the coefficient α is used as in the above [Equation 8]. By using the coefficient α, the third predetermined value H _max2 and the fourth predetermined value H _min2 are determined. In the lower part of FIG. 11, the magnitudes of the arrows indicated by D (i2) and D (i3) are differential luminance values, and samples i2 and i3 taking these differential luminance values D (i2) and D (i3) are blood. Two boundary points of the column reflection region.

  When the process of step S420 to step S470 is completed for the start point side SH, the measurement range is changed to the end point side EH, and the process of step S420 to step S470 is similarly executed for the end point side EH. The two end points of the intersecting blood flow regions and the two boundary points of the blood column reflection region are specified.

  FIG. 12 is a diagram illustrating the relationship between the luminance value and the differential luminance value when a part of the blood vessels on both sides is included in the measurement range. As shown in the upper part of FIG. 12, when the central blood vessel among the three blood vessels is completely included in the measurement range and a part of the blood vessels on both sides is included in the measurement range, the maximum value near both ends of the measurement range, Minimum values may appear. However, as shown in [Formula 6] above, in step S440, the search range of the maximum value is a sample from i = 0 to i = s / 2 (the left half in FIG. 12). The maximum value appearing near = s is not searched. Similarly, since the search range of the minimum value is a sample from i = s / 2 to i = s (right half in FIG. 12), the minimum value that appears in the vicinity of sample i = 0 is not searched. Therefore, even when a blood vessel other than a blood vessel that the user wants to measure is included near both ends of the measurement range, the blood flow region 2 end point of the intended central blood vessel without being influenced by the blood vessel, blood Two boundary points of the column reflection region can be specified.

  When the two end points of the blood flow region and the two boundary points of the blood column reflection region are specified for both the start point side SH and the end point side EH and the process of step S400 is completed, the blood column reflection width ratio calculating unit 60 is then performed. However, the blood column reflection width ratio, which is a kind of blood flow width ratio, is calculated (step S500). Specifically, the blood column reflection width ratio is calculated by executing processing according to the following [Equation 10].

[Formula 10]
Blood column reflection width ratio = (total value of distance between two boundary points of two blood column reflection areas calculated based on each measurement graphic) / (two blood flows calculated based on each measurement graphic) Sum of distances between two end points of the area)

  As shown in the above [Equation 10], the blood column reflection width ratio is (the total value of the distance between two boundary points of two blood column reflection regions calculated based on each measurement graphic) (for each measurement). It is obtained by dividing by the sum of the distances between the two end points of the two blood flow regions calculated based on the figure. “Distance between two boundary points” means a distance between boundary points obtained based on the same side (measurement range), and “Distance between two end points” means the same side (measurement range) Means the distance between the end points obtained based on. When only one measurement graphic is arranged, the sum of the two distances of the distance based on the start point side and the distance on the end point side is the total value. Since the total value is divided by the total value, it can be said that the ratio of each distance is averaged. That is, when only one measurement figure is arranged, the blood column reflection width ratio is obtained as an average of the distance at the start point side and the distance at the end point side by [Equation 10]. In the present invention, the blood column reflection width ratio can also be calculated using only the start point side or the end point side. However, with only one side, if there is spot-like luminance unevenness on either the start point side or the end point side, an appropriate blood column reflection width ratio may not be obtained. Therefore, in order to alleviate the influence of spot-like luminance unevenness and obtain an appropriate blood column reflection width ratio, an average of the start side and the end side is obtained. Even when only the start point or the end point is used, when a plurality of measurement figures are arranged, the influence of spot-like luminance unevenness can be reduced by obtaining an average of them. The blood column reflection width ratio calculating unit 60 stores the calculated blood column reflection width ratio in the information storage unit 80. In parallel, the display means 20 displays the calculated blood column reflection width ratio.

  After the calculated blood column reflection width ratio is calculated, a graphic based on the measurement result is displayed (step S600). Specifically, a trapezoid connecting the two end points of the blood flow region specified based on the start point side and the two end points of the blood flow region specified based on the end point side is connected to the midpoint of the two pairs of two end points. A measuring figure consisting of a trapezoid connecting a line segment and two boundary points of the blood column reflection region specified based on the start point side and two boundary points of the blood column reflection region specified based on the end point side is displayed on the display means 20. indicate. Specifically, first, using the two end points i1 and i4 of the blood flow region specified in step S460, for the start point side, X11 = X (i1), Y11 = Y (i1), X14 = X ( i4), Y14 = Y (i4), and the end points are X21 = X (i1), Y21 = Y (i1), X24 = X (i4), Y24 = Y (i4) Identify the corresponding 4 vertices. Subsequently, using the two boundary points i2 and i3 of the blood column reflection region specified in step S470, X12 = X (i2), Y12 = Y (i2), X13 = X (i3) for the start point side. , Y13 = Y (i3), and the end points are X22 = X (i2), Y22 = Y (i2), X23 = X (i3), and Y23 = Y (i3). 4 vertices corresponding to are specified.

  Next, the display means 20 includes the four end points (X11, Y11), (X14, Y14), (X21, Y21), (X24, Y24) of the specified blood flow region and the specified blood column reflection region. 4 boundary points (X12, Y12), (X13, Y13), (X22, Y22), (X23, Y23), and the midpoint (X1, Y1) of the two end points on the start side (= X11 + X14) / 2 (Y11 + Y14) / 2), the midpoint (X2, Y2) = ((X21 + X24) / 2, (Y21 + Y24) / 2) of the two end points on the end point side, and a trapezoid connecting four vertices corresponding to the four end points, respectively. Then, the line segment connecting the middle point on the start point side and the end point side and the trapezoid connecting the four vertices respectively corresponding to the four boundary points are replaced with the measurement graphic already arranged and displayed. As a result, the user can grasp which range is specified as a blood flow region and which range is specified as a blood column reflection region by a graphic based on the displayed measurement result. However, as described in step S500, in calculating the blood column reflection width ratio, the measurement results on the start and end sides of a single measurement graphic are averaged. In other words, the calculation results on the start and end sides are replaced with the average values of both, and corrected so that both are the same. The process of converting from trapezoid to rectangle is also performed. Specifically, the larger of the distance between the end point (X11, Y11) on the start point side and the midpoint (X1, Y1) and the distance between the end point (X14, Y14) and the midpoint (X1, Y1) is greater. Let WO1 be the distance between the end point (X21, Y21) on the end point side and the midpoint (X2, Y2) and the distance between the end point (X24, Y24) and the midpoint (X2, Y2), whichever is greater is WO2 Then, the four end points are corrected so that WO1 and WO2 are average values (WO1 + WO2) / 2 = WO, and the distance between the starting point side boundary point (X12, Y12) and the middle point (X1, Y1) and the boundary point WI1 is the greater of the distances between (X13, Y13) and the midpoint (X1, Y1), and the distance between the border point (X22, Y22) and the midpoint (X2, Y2) on the end side and the border point The greater of the distance between (X23, Y23) and the midpoint (X2, Y2) When WI2 towards corrects the four boundary points to WI1 and WI2 is the average value (WI1 + WI2) / 2 = WI. FIG. 9B shows an example of a figure displayed in accordance with the identified blood flow region and blood column reflection region after such correction. The shaded portion is a portion corresponding to a region other than the blood column reflection region in the blood flow region of the fundus image displayed in an overlapping manner. WI / WO obtained by dividing the width WI shown in FIG. 9B by the width WO corresponds to the blood column reflection width ratio calculated by [Formula 10]. When the two end points of the blood flow region and the two boundary points of the blood column reflection region are specified in step S400, as described above, the blood flow region specified corresponding to the two sides and corrected to the rectangle in step S600. A rectangle that includes two sides that connect the two end points, a line segment that connects the midpoint of the two end points on the start point side and the midpoint of the two end points on the end point side, and is corrected to a rectangle specified for the two sides A graphic consisting of a rectangle including two sides connecting the two boundary points of the blood column reflection region is displayed.

  This figure corresponds to the rectangle that connects the four endpoints of the blood flow area specified by the calculation, the line that connects the midpoint of the two end points on the start point side and the midpoint of the two end points on the end point side, and two sides A rectangle including two sides connecting the two end points of the blood column reflection area specified in this manner is displayed. However, the measurement process ends in an instant from the human sense, so after placing the measurement figure given by a single rectangle, when the execution of the measurement process is instructed, the blood flow area and the blood column reflection area are instantaneously It looks like it has changed to a double rectangular measurement figure.

  The processes in steps S500 and S600 are not necessarily performed, and any of them may be performed first. It is also possible for the blood flow region end point specifying means 50 to directly store the information of the end points of the blood flow region and the boundary points of the blood column reflection region obtained in step S400 in the information storage unit 80 and finish the processing. is there.

  FIG. 13 is a diagram illustrating an example in which a processed graphic is displayed superimposed on an artery in the analysis range of an actual fundus image. Corresponding to the measurement graphic arranged in the state shown in FIG. 7A, as shown in FIG. 13, a graphic corresponding to the width of the blood flow region and the width of the blood column reflection region is displayed. Accordingly, the rectangular measurement graphic shown in FIG. 7A may be fitted (positioned) to each of the blood flow region and the blood column reflection region, and may appear to be separated and deformed into a double rectangle. . In the case of a normal blood vessel in which no blood column reflex occurs at all (blood column reflex width ratio: 0.0), or in the case of severe arteriosclerosis in which the entire blood flow region exhibits a blood column reflex (respiratory column width ratio: 1. In 0), the rectangle is fitted to the blood flow region and does not separate twice. Therefore, in the case of severe arteriosclerosis, measurement is impossible, but since the color of the blood vessel changes to copper or gold, it can be distinguished from a normal blood vessel.

  FIG. 14 is a diagram showing an example in which measurement figures are arranged at three locations in the analysis range. 14A shows an image of the analysis range, and FIG. 14B shows a calculation result of the blood column reflection width ratio (blood flow width ratio) based on a plurality of measurement figures. A blood column reflection width ratio of 0.5 or less is normal. In the example of FIG. 14, the blood column reflection width ratio is 0.23, which is normal.

  FIG. 15 is a diagram illustrating an example in which measurement graphics are arranged at three locations in the analysis range. 15A shows the entire fundus image, FIG. 15B shows the image in the analysis range, and FIG. 15C shows the blood column reflection width ratio (based on a plurality of measurement figures). The calculation result of blood flow width ratio is shown. In the example of FIG. 15, the blood column reflection width ratio is 0.51 and exceeds 0.5, so that arteriosclerosis is progressing.

<3. Functions as a graphic display device for measurement>
Although the fundus image processing apparatus has been described in the above embodiment, the fundus image processing apparatus also functions as a measurement graphic display device that is a display device that displays a measurement graphic. That is, when the end point of the blood flow region of the blood vessel included in the fundus image is specified, the outer shape is a trapezoid (concept including a rectangle as one aspect) based on an instruction from the instruction input unit 10, and two parallel sides A measurement graphic having a reference line connecting the midpoints is displayed on the display means 20. Then, based on an instruction from the instruction input unit 10, the measurement range setting unit 40 sets each of two parallel sides of the arranged measurement graphic as a measurement range. Further, the blood flow region end point specifying means 50 specifies the two end points of the blood flow region and the two boundary points of the blood column reflection region based on each of the two parallel sides of the set measurement range. A graphic obtained by combining a trapezoid having four apexes at the four end points of the identified blood flow region and a trapezoid having the four boundary points of the identified blood column reflection region as apexes is displayed on the display unit 20 instead of the measurement graphic. . That is, the fundus image processing apparatus shown in FIGS. 1 and 2 can function as a measurement graphic display apparatus. In this case, the measurement figure is deformed like a so-called measurement gauge and gives an impression as if fitting to the blood flow region and the blood column reflection region, and the blood flow width and the blood column reflection width in the blood flow region are It becomes easy to grasp intuitively. In the above embodiment, the measurement graphic is given as a rectangle, and the external shape of the graphic calculated after the measurement process is corrected from the trapezoid to the rectangle and displayed. However, if the start point side and the end point side are parallel, other shapes are displayed. The opposing sides may not be parallel. That is, it may be a trapezoid that is not rectangular.

<4. Modified example>
The preferred embodiments of the present invention have been described above. However, the present invention is not limited to the above embodiments, and various modifications can be made. For example, in the above embodiment, a full-color image is used as a fundus image obtained by photographing. However, a monochrome image photographed by a digital fundus camera using an infrared light source is also being used, and in this case, the fundus blood vessel is clearly contrasted. Therefore, the binarization process in the process of specifying the blood flow region can be easily performed with high accuracy by a simple discriminant analysis method. In recent years, the gradation of commercial digital cameras has been expanded to 10 bits or more, and it is also possible to acquire full-color images such as 1024 gradations for each color, using color images with a larger number of gradations. May be.

1 ... CPU (Central Processing Unit)
2 ... RAM (Random Access Memory)
3. Storage device 4. Touch panel input I / F
5. Data input / output I / F
DESCRIPTION OF SYMBOLS 6 ... Display part 7 ... Image input part 10 ... Instruction input means 20 ... Display means 30 ... Image expansion means 40 ... Measurement range setting means 50 ... Blood flow area | region end point specification Means 60: Blood column reflection width ratio calculating means 70 ... Fundus image storage means 80 ... Information storage means 100 ... Fundus image processing apparatus

Claims (10)

An apparatus for processing a fundus image,
Display means for displaying the fundus image and a measurement graphic having a trapezoidal outer shape;
Based on an instruction from the outside, a measurement range setting unit that sets at least one side of two parallel sides of the measurement graphic as a measurement range based on the measurement graphic arranged to overlap the fundus image;
Sampling the sides set as the measurement range, creating a one-dimensional luminance array that is an array of luminance values corresponding to each sample obtained by sampling,
Create a differential array that is an array of differential luminance values using the difference value with the luminance value of the adjacent sample using the luminance array,
Search for the maximum sample taking the maximum value and the minimum sample taking the minimum value for the differential array,
A second minimum sample having a minimum value in the range from the maximum sample to the midpoint is searched for the differential array, and a second maximum sample having a maximum value in the range from the minimum sample to the midpoint is searched. ,
Identifying two end points of the blood flow region based on the maximum sample and the minimum sample, and specifying two boundary points of the blood column reflection region based on the second maximum sample and the second minimum sample Means,
A fundus image processing apparatus comprising:   Based on the ratio between the distance between the two boundary points of the two blood column reflection regions specified based on the measurement graphic and the distance between the two end points of the two blood flow regions specified based on the measurement graphic The fundus image processing apparatus according to claim 1, further comprising blood column reflection width ratio calculating means for calculating a blood column reflection width ratio.   The image processing apparatus further includes an image enlarging unit that trims a designated area of the fundus image based on designation from the outside and performs an enlargement process of the fundus image so that the area has a predetermined number of pixels. The fundus image processing apparatus according to claim 1 or 2. The blood flow region end point specifying means includes:
When creating the luminance array, if the fundus image is a color image in which each component of RGB has a predetermined gradation, the gradation value of the G component and the gradation value of the B component are inverted with respect to the fundus image. The fundus image processing according to any one of claims 1 to 3, wherein the luminance value is converted to a grayscale image by a square root of a product of the values and further inverted as a negative value. apparatus.   The fundus image processing apparatus according to any one of claims 1 to 4, wherein the measurement graphic includes a reference line that connects midpoints of two parallel sides of the trapezoid. The blood flow region end point specifying means includes:
When creating a luminance array,
Sample one of two parallel sides of the measurement graphic, specify two points where each sample is moved by a predetermined distance in two directions orthogonal to the side, and brightness at the two specified points The fundus image processing apparatus according to any one of claims 1 to 5, wherein a luminance value corresponding to each sample is corrected using a value. The blood flow region end point specifying means includes:
A sample having a differential luminance value smaller than the first predetermined value near the first sample when viewed from the maximum point where the differential luminance value is smaller than the maximum sample and the differential luminance value is larger than the first predetermined value is the first blood flow region. Identified as an endpoint,
A blood sample having a differential luminance value larger than a second predetermined value near the last sample when viewed from a minimum point having a differential luminance value larger than the minimum sample and smaller than a second predetermined value smaller than a first predetermined value. Identified as the second endpoint of the flow region,
A sample having a differential luminance value smaller than the third predetermined value near the top sample as seen from the maximum point where the differential luminance value is smaller than the second maximum sample and the differential luminance value is larger than the third predetermined value is the blood column reflection region. A fourth predetermined value that is specified as a first boundary point and is closer to the last sample as viewed from a minimum point having a differential luminance value larger than the second minimum sample and smaller than a fourth predetermined value that is smaller than a third predetermined value. The fundus image processing apparatus according to any one of claims 1 to 6, wherein a sample having a larger differential luminance value is specified as the second boundary point of the blood column reflection region.   The display means includes a trapezoid whose apex is the four end points of the blood flow region specified on each of the two parallel sides, and four boundary points of the blood column reflection region specified on each of the two parallel sides. The fundus image processing apparatus according to any one of claims 1 to 7, wherein a figure obtained by combining a trapezoid with a vertex as a top is displayed so as to overlap the fundus image.   A program for causing a computer to function as the fundus image processing apparatus according to any one of claims 1 to 8. A display device for displaying a measurement graphic when specifying a blood flow region of a blood vessel included in a fundus image,
Based on an instruction from the outside, the external shape is a trapezoid, and a measurement figure with a reference line connecting the midpoints of two parallel sides is displayed.
Two parallel sides of the graphic for measurement arranged based on an instruction from the outside are set as measurement ranges, and two boundaries of the blood flow region and two boundaries of the blood column reflection region are set based on the two sides of the set measurement range. A figure obtained by combining a trapezoid having four apexes with four end points of the specified blood flow region and a trapezoid having four apex points with the four boundary points of the specified blood column reflection region is identified. A graphic display device for measurement, characterized by being displayed instead of.