CN109767396B - Oral cavity CBCT image denoising method based on image dynamic segmentation - Google Patents
Oral cavity CBCT image denoising method based on image dynamic segmentation Download PDFInfo
- Publication number
- CN109767396B CN109767396B CN201910006387.4A CN201910006387A CN109767396B CN 109767396 B CN109767396 B CN 109767396B CN 201910006387 A CN201910006387 A CN 201910006387A CN 109767396 B CN109767396 B CN 109767396B
- Authority
- CN
- China
- Prior art keywords
- image
- images
- segmentation
- pixel
- foreground
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000011218 segmentation Effects 0.000 title claims abstract description 30
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000007408 cone-beam computed tomography Methods 0.000 title claims abstract 7
- 210000000214 mouth Anatomy 0.000 title description 3
- 238000012545 processing Methods 0.000 claims abstract description 15
- 230000009467 reduction Effects 0.000 claims abstract description 15
- 238000001914 filtration Methods 0.000 claims abstract description 11
- 230000002146 bilateral effect Effects 0.000 claims abstract description 8
- 210000004872 soft tissue Anatomy 0.000 claims description 4
- 238000003709 image segmentation Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 description 4
- 210000000845 cartilage Anatomy 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 2
- 210000000988 bone and bone Anatomy 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003706 image smoothing Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000010410 layer Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 210000003625 skull Anatomy 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000003325 tomography Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
An oral CBCT image denoising method based on image dynamic segmentation selects a group of reconstructed CBCT slice images and selects an image to be denoised from the images; selecting 2N images adjacent to the left and right of an image to be denoised, wherein N represents the number of the images selected to the left or the right, and carrying out median filtering on the 2N +1 images to obtain a filtered image in order to weaken the influence of noise on the segmentation of the subsequent image; a pixel threshold is given, the pixel threshold is compared with a pixel value at any point in the image, the value is assigned again, and an image with segmented threshold is obtained by traversing each pixel point in the image; the deviation of these pixel values is calculated: calculating the deviation of all pixel points and adjacent images to obtain a deviation image, and giving another pixel threshold value to obtain another threshold value segmentation image; performing OR operation on the two segmented images to obtain a binary image, so that the foreground part and the background part of the image can be marked; and performing noise reduction processing on the foreground and the background of the image to different degrees by adopting a bilateral denoising algorithm to obtain a denoised image.
Description
Technical Field
The invention relates to the technical field of radiation imaging, in particular to an oral CBCT image denoising method based on image dynamic segmentation.
Background
Conebeam CT has the characteristics of good real-time property, high sensitivity, convenient use and the like, and is widely applied to clinical diagnosis and treatment of oral cavities and skull parts. However, due to the influence of factors such as air scattering and electronic noise in image acquisition, a large amount of noise exists in the CBCT image, and the noise seriously affects the quality of the image, thereby affecting the observation and judgment of a doctor on a focus. Therefore, it is a subject of constant research by students in various fields how to find a suitable CBCT image denoising method, which can reduce image noise and keep the details and edge information of the image as much as possible.
When the noise of the image is reduced, the projection image data, the image data in the reconstruction process or the reconstructed image data can be processed selectively; at present, there are many CBCT image denoising methods, and the traditional methods, such as mean, median, gaussian denoising, etc., lose the details of the image while removing the noise, so that the edge of the image becomes blurred, and the quality of the image is reduced; in recent years, applications such as TV noise reduction, non-local mean denoising, partial differential denoising, and the like are widely applied, but these methods all use an entire image as a processing unit, and perform denoising processing based on redundancy of the image itself, and although the methods are improved compared with the conventional methods, the effects and the detail retentivity are limited. In fact, for the oral CT, when a doctor diagnoses, the detailed information of the teeth and the tooth root parts is a major concern, and the information displayed in each layer of the slices is changed; in the soft tissue part, the information displayed by the front and back slices is not changed too much, and the part is not concerned too much. Therefore, the method can be distinguished for processing in denoising; for parts with important attention, details are kept as much as possible; for less interesting parts, the filtering can be increased to obtain a visual optimum.
At present, no algorithm for denoising a CBCT image after segmenting the image is provided.
Disclosure of Invention
The invention solves the problems: the oral CBCT image denoising method based on image dynamic segmentation can effectively segment a concerned foreground region and a background region, and then perform denoising to the segmented foreground and background images in different degrees respectively to obtain an optimal denoising image.
The technical scheme adopted by the invention is as follows: a method for denoising an oral CBCT image based on image dynamic segmentation is disclosed, the processing flow of which is shown in FIG. 1, and the method comprises the following steps:
the method comprises the following steps: selecting a group of reconstructed CBCT slice images, and selecting an image f to be denoised from the CBCT slice imagesi(i denotes the second in the image sequencei-frame);
step two: selecting and denoising image fi2N images (N represents the number of images selected leftwards or rightwards) adjacent to each other are subjected to median filtering to reduce the influence of noise on the segmentation of the subsequent images, and a filtered image f is obtainedi-N′…fi′…fi+N′；
Step three: given a pixel threshold T1, for image fi' Pixel value f at any point (x, y)i' (x, y) is compared with T1, and is reassigned to g (x, y) according to formula (1), and a threshold segmentation image is obtained by traversing each pixel point in the image, which is called g 1;
step four: selecting any coordinate (x, y) in the image to obtain an image fi-N′…fi′…fi+N' pixel value f corresponding at this coordinatei-N′(x，y)…fi′(x，y)…fi+N' (x, y). Calculating the pixel values of fi' (x, y) deviation σ (x, y):
step five: calculating fiObtaining a deviation image by the deviation of all pixel points and adjacent images, giving another pixel threshold value T2, and performing threshold value segmentation operation according to the third step to obtain another threshold value segmentation image g2, wherein the segmentation can segment out soft tissues and tooth roots with changed left and right frames in the image;
step six: performing OR operation on the two segmented images g1 and g2 obtained in the third step and the fifth step to obtain a binary image g with a foreground of 1 and a background of 0;
step seven: combining the binary image g in the step six, and pairing the image f to be denoisediThe foreground and the background of the image are subjected to noise reduction processing of different degrees by adopting a bilateral noise reduction algorithm respectively to obtain a noise-reduced image. The specific method for performing noise reduction processing on the image after image segmentation comprises the following steps: calibrating the foreground and the background of the image by 1 and 0 respectively through threshold segmentation, and then processing the image f to be denoisediThe foreground and the background of the image are subjected to noise reduction processing of different degrees by adopting a bilateral noise reduction algorithm respectively to obtain a noise-reduced image.
Compared with the prior art, the invention has the advantages that:
at present, the existing CBCT image denoising methods are all processed on the basis of a whole image, and a contradiction exists between detail maintenance and image smoothing in the processing process. The oral CBCT image denoising method based on image dynamic segmentation provided by the invention can segment the image part focused by a doctor and the image part not focused by the doctor, and carry out denoising treatment respectively, and maintain the details of the focused part as much as possible; for less interesting parts, the filtering can be increased to obtain a visual optimum. The innovation points of the invention are mainly two:
(1) innovation of the segmentation method: if the single-layer image is segmented according to the pixel value and a judgment factor, parts such as tooth cracks, tooth roots, cartilages and the like are likely to serve as backgrounds, and therefore the invention provides an image segmentation algorithm based on dynamic change. The algorithm treats a set of CT slice images as a motion picture, for which the static part can be considered as the background and for which the dynamic part can be considered as the foreground. According to the characteristic that the background images before and after each slice of the CT image have similarity and the foreground part is changed, firstly, selecting an image to be denoised, and selecting N frames of adjacent slice images to the left and the right of the image to be denoised respectively; then, the change of each point on the image in the time direction is obtained, the value of the point with the change before and after is larger than that of the point without the change too much, and therefore the changed foreground point can be marked. The segmentation can segment cartilage and tooth root parts with changed left and right frames in the image;
(2) innovation of the image noise reduction method: the new denoising thought provided by the invention firstly segments the image, separates the foreground region concerned by doctors and the background region not concerned by doctors for denoising, and can store detail information as much as possible for the foreground region and enhance filtering and smoothen the background region as much as possible, thereby obtaining an optimal denoising image.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a diagram of an image to be denoised in the present invention;
FIG. 3 is a segmented image g1 in accordance with the present invention;
FIG. 4 is a segmented image g2 in accordance with the present invention;
FIG. 5 is a binary image g in the present invention;
FIG. 6 is a diagram of the denoising result in the present invention;
fig. 7 is a diagram of a noise reduction result of bilateral filtering.
Detailed Description
The invention is described in further detail below with reference to the accompanying examples.
As shown in fig. 1, the method of the present invention is specifically implemented as follows:
the method comprises the following steps: selecting a group of reconstructed CBCT (cone beam projection computer reconstruction tomography) slice images, and selecting an image f to be denoised from the slice imagesi(ii) a As shown in fig. 2, is the ith image of a set of CBCT axial plane slice images.
Step two: selecting and denoising image fi2N images (N represents the number of images selected leftwards or rightwards) adjacent to each other are subjected to median filtering to reduce the influence of noise on the segmentation of the subsequent images, and a filtered image f is obtainedi-N′…fi′…fi+N′；
Step three: given a pixel threshold T1 (1800 in this embodiment of the invention, T1) for image f, the size of which is related to the range of pixel values of the imagei' Pixel value f at any point (x, y)i' (x, y) is compared with T1 and reassigned to g (x, y) according to equation (1), and a threshold value can be obtained by traversing each pixel point in the imageThe segmented image, referred to as g1, has white portions as the foreground of segmentation as shown in fig. 3, which can segment teeth and hard bone portions with relatively large pixel values;
step four: selecting any coordinate (x, y) in the image to obtain an image fi-N′…fi′…fi+N' pixel value f corresponding at this coordinatei-N′(x，y)…fi′(x，y)…fi+N' (x, y). Calculating the pixel values of fi' (x, y) deviation σ (x, y):
step five: calculating fi' the deviation of all the pixels from the neighboring image is used to obtain a deviation image σ, then another pixel threshold T2 (T2 is 176 in the embodiment of the present invention, whose size is related to the range of the deviation value obtained in step 4) is given, and a threshold segmentation operation is performed according to step three to obtain another threshold segmentation image g 2. The segmentation can segment cartilage and root parts with some left and right frames changed in the image, as shown in the white part of fig. 4;
step six: performing an or operation on the two segmented images g1 and g2 obtained in the third step and the fifth step to obtain a binary image g with a foreground of 1 and a background of 0, wherein the binary image is a relatively comprehensive foreground image; as shown in fig. 5, the white part of the image has a pixel value of 1, and is a separated foreground image part; the black portion has a pixel value of 0 and is the background portion of the image.
Step seven: combining the binary image g in the step six, and processing the image f to be denoisediThe foreground and the background of the image are respectively subjected to noise reduction processing of different degrees by adopting a bilateral denoising algorithm:
wherein f isi(x) For inputting an image, hi(x) For the filtered image, c (θ, x) and s (f)i(θ)，fi(x) Respectively, a domain kernel function and a value domain kernel function, and k (x) is a normalization parameter, which are calculated by the following formula:
by adjusting the parameter σdAnd σrThe value of (c) can control the intensity of the filtering, and σ is the greater the intensity of the filtering for the background portiondAnd σrThe set value is larger (the experimental value is sigma)d＝17，σr230); for the foreground part σdAnd σrThe set value is small (the implementation of the invention takes the full value as sigma)d＝13，σr176). The noise reduction result of fig. 1 is shown in fig. 6, and the result is that the detail information of the tooth, the tooth root and the like is saved as much as possible, and the soft tissue part is smoothed. Fig. 7 is a process of bilateral filtering directly on the image, and it can be seen that the loss of edge information of the image is serious, and the loss of details of the root and fissure parts is also serious.
The above examples are provided only for the purpose of describing the present invention, and are not intended to limit the scope of the present invention. The scope of the invention is defined by the appended claims. Various equivalent substitutions and modifications can be made without departing from the spirit and principles of the invention, and are intended to be within the scope of the invention.
Claims (2)
1. An oral CBCT image denoising method based on image dynamic segmentation is characterized by comprising the following steps:
the method comprises the following steps: selecting a group of reconstructed CBCT slice images, and selecting an image f to be denoised from the CBCT slice imagesiI denotes the ith frame in the image sequence;
step two: selecting and denoising image fi2N images adjacent to each other left and right, wherein N represents the number of the images selected left or right, and in order to weaken the influence of noise on the segmentation of the subsequent images, the 2N +1 images are subjected to median filtering to obtain a filtered image fi-N′…fi′…fi+N′；
Step three: given a pixel threshold T1, for image fi' A pixel value f at any point (x, y) ini' (x, y) is compared to T1 and operated according to equation (1) to compare fi' assigning value to g (x, y) again, and traversing each pixel point in the image to obtain a threshold segmentation image, which is called g 1;
step four: selecting any coordinate (x, y) in the image to obtain an image fi-N′…fi′…fi+N' pixel value f corresponding at this coordinatei-N′(x，y)…fi′(x，y)…fi+N' (x, y) calculating pixel values of fi' (x, y) deviation σ (x, y):
step five: calculating fi' deviation of all pixel points in the three-dimensional image and adjacent images obtains a deviation image, and gives another pixel threshold value T2, and performs threshold segmentation operation according to the third step to obtain another threshold segmentation image g2,the segmentation can segment the soft tissue and the tooth root part with the changed left and right frames in the image;
step six: performing OR operation on the two segmented images g1 and g2 obtained in the third step and the fifth step to obtain a binary image g with a foreground of 1 and a background of 0;
step seven: combining the binary image g in the step six, and pairing the image f to be denoisediThe foreground and the background of the image are subjected to noise reduction processing of different degrees by adopting a bilateral noise reduction algorithm respectively to obtain a noise-reduced image.
2. The oral CBCT image denoising method based on image dynamic segmentation as claimed in claim 1, wherein in the seventh step, the specific method for denoising the image after image segmentation is as follows:
calibrating the foreground and the background of the image by 1 and 0 respectively through threshold segmentation, and then processing the image f to be denoisediThe foreground and the background of the image are subjected to noise reduction processing of different degrees by adopting a bilateral noise reduction algorithm respectively to obtain a noise-reduced image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910006387.4A CN109767396B (en) | 2019-01-04 | 2019-01-04 | Oral cavity CBCT image denoising method based on image dynamic segmentation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910006387.4A CN109767396B (en) | 2019-01-04 | 2019-01-04 | Oral cavity CBCT image denoising method based on image dynamic segmentation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109767396A CN109767396A (en) | 2019-05-17 |
| CN109767396B true CN109767396B (en) | 2021-04-02 |
Family
ID=66453537
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910006387.4A Active CN109767396B (en) | 2019-01-04 | 2019-01-04 | Oral cavity CBCT image denoising method based on image dynamic segmentation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109767396B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111008949B (en) * | 2019-08-16 | 2021-09-14 | 苏州喆安医疗科技有限公司 | Soft and hard tissue detection method for tooth image |
| CN111862113A (en) * | 2020-07-09 | 2020-10-30 | 常州飞诺医疗技术有限公司 | A method for extracting root and alveolar bone based on CBCT image segmentation |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101127117B (en) * | 2007-09-11 | 2010-05-26 | 华中科技大学 | A Method for Segmenting Vascular Data Using Sequential Digital Subtraction Angiography Images |
| CN102014240B (en) * | 2010-12-01 | 2013-07-31 | 深圳市蓝韵实业有限公司 | Real-time medical video image denoising method |
| CN103150712B (en) * | 2013-01-18 | 2016-04-27 | 清华大学 | Image denoising method based on projection sequence data similarity |
| WO2016011489A1 (en) * | 2014-07-23 | 2016-01-28 | The University Of Sydney | Thoracic imaging for cone beam computed tomography |
| CN105184741A (en) * | 2015-08-03 | 2015-12-23 | 山东师范大学 | Three-dimensional CBCT (cone-beam computed tomography) image denoising method on the basis of improved nonlocal means |
| CN105761252B (en) * | 2016-02-02 | 2017-03-29 | 北京正齐口腔医疗技术有限公司 | The method and device of image segmentation |
| CN106846359B (en) * | 2017-01-17 | 2019-09-20 | 湖南优象科技有限公司 | Moving target rapid detection method based on video sequence |
| CN108932716B (en) * | 2017-05-26 | 2020-09-22 | 无锡时代天使医疗器械科技有限公司 | Image segmentation method for dental images |
| CN108876769B (en) * | 2018-05-31 | 2020-11-03 | 厦门大学 | A method for segmentation of left atrial appendage CT images |
-
2019
- 2019-01-04 CN CN201910006387.4A patent/CN109767396B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109767396A (en) | 2019-05-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107808156B (en) | Region of Interest Extraction Method | |
| WO2018120644A1 (en) | Blood vessel extraction method and system | |
| CN106909947B (en) | Mean Shift algorithm-based CT image metal artifact elimination method and system | |
| CN115375574B (en) | Multi-scale non-local low-dose CT image denoising method based on region adaptation | |
| CN109767396B (en) | Oral cavity CBCT image denoising method based on image dynamic segmentation | |
| Vyavahare et al. | Segmentation using region growing algorithm based on CLAHE for medical images | |
| CN114299081B (en) | Maxillary sinus CBCT image segmentation method, maxillary sinus CBCT image segmentation device, maxillary sinus CBCT storage medium and electronic equipment | |
| Kulathilake et al. | [Retracted] InNetGAN: Inception Network‐Based Generative Adversarial Network for Denoising Low‐Dose Computed Tomography | |
| CN113936068B (en) | Artifact correction method, device and storage medium | |
| CN113706687A (en) | Nose environment modeling method and device for path planning | |
| Idowu et al. | Improved enhancement technique for medical image processing | |
| CN113643393A (en) | A Metal Artifact Correction Method for CBCT Images Based on Guided Graph Filtering | |
| CN114663653B (en) | Window level window width calculation method for medical image region of interest | |
| CN111091514B (en) | Oral cavity CBCT image denoising method and system | |
| Zhang et al. | Metal artifact reduction based on the combined prior image | |
| CN105701860A (en) | Volume rendering method | |
| CN117036311A (en) | Mammary gland segmentation method, system and medium based on progressive growth learning | |
| CN117059232A (en) | Registration method based on multimode image data volume space | |
| CN114022879B (en) | Squamous cell structure enhancement method based on optical fiber microscopy endoscope image | |
| Nur‘Aini et al. | Comparative Analysis of Image Filtering for Dental Caries Image Denoising | |
| CN108171768A (en) | A kind of low dosage CBCT image rebuilding methods based on BM3D | |
| Park et al. | Performance evaluation of improved median-modified Wiener filter with segmentation method to improve resolution in computed tomographic images | |
| Guo et al. | A Low-Dose CT Image Denoising Method Combining Multistage Network and Edge Protection | |
| CN114418894B (en) | Oral cone beam CT 3D image denoising method and system based on anatomical priors | |
| Vidyasaraswathi et al. | Gradient, texture driven based dynamic-histogram equalization for medical image enhancement |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 100084 A800B, 8 floor, Tsinghua Tongfang mansion, Tsinghua Yuan, Haidian District, Beijing Patentee after: Beijing Langshi Instrument Co.,Ltd. Address before: 100084, Beijing Haidian District Tsinghua Yuan, Tsinghua Tongfang building, 8 floor, A800B Patentee before: LARGEV INSTRUMENT Corp.,Ltd. |