CN111259889B - Image text recognition method, device, computer equipment and computer storage medium - Google Patents
Image text recognition method, device, computer equipment and computer storage medium Download PDFInfo
- Publication number
- CN111259889B CN111259889B CN202010051370.3A CN202010051370A CN111259889B CN 111259889 B CN111259889 B CN 111259889B CN 202010051370 A CN202010051370 A CN 202010051370A CN 111259889 B CN111259889 B CN 111259889B
- Authority
- CN
- China
- Prior art keywords
- text
- image
- region
- information
- text region
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/146—Aligning or centring of the image pick-up or image-field
- G06V30/1475—Inclination or skew detection or correction of characters or of image to be recognised
- G06V30/1478—Inclination or skew detection or correction of characters or of image to be recognised of characters or characters lines
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The application discloses an image text recognition method, an image text recognition device and a computer storage medium, relates to the technical field of computers, and aims at an image with a complex scene, so that a text recognition result can be structured, and the accuracy of image text recognition is improved. The method comprises the steps of obtaining an image to be recognized, preprocessing the image to be recognized to obtain a target recognition image, determining position information of a text region in the target recognition image based on a pre-trained text region detection model, inputting the target recognition image and the position information of the text region in the target recognition image into the pre-trained text recognition model to obtain text information in the text region, and carrying out structuring processing on the text information in the text region to obtain a text field with a mapping relation.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to an image text recognition method, an image text recognition device, a computer device, and a computer storage medium.
Background
With the development of technology, images play a great role in information dissemination. To better promote the advertising effect, more and more images are added with text, for example, in the medical transaction platform, a medical institution needs a user to upload an invoice image, so that the invoice can be checked based on the text content in the uploaded invoice image. Therefore, the text in the image usually contains rich information, so that the text in the image is extracted and identified, and the method has important significance in the aspects of analysis, understanding, information retrieval and the like of the image content.
The existing image text recognition method generally detects text information boxes in images, then recognizes the detected text information boxes, and finally returns a recognition result, so that automatic recognition is achieved, and labor input cost is saved.
However, in the image of the actual application scene, there are complicated and various image contents, invoice images or texts in many natural images, which are generally affected by the background contents of irregular images, so that the existing image text recognition method has many conditions of missed detection and false detection, the recognition accuracy is low, the final text recognition result is incomplete, the fields obtained by recognition cannot correspond to the required fields, and the subsequent text use is seriously affected.
Disclosure of Invention
In view of the above, the present invention provides an image text recognition method, apparatus, computer device and computer storage medium, which mainly aims to solve the problem of low accuracy of image text recognition for complex scenes at present.
According to one aspect of the present invention, there is provided an image text recognition method, the method comprising:
Acquiring an image to be identified, and preprocessing the image to be identified to obtain a target identification image;
Determining the position information of a text region in the target recognition image based on a pre-trained text region detection model;
Inputting the target recognition image and the position information of the text region in the target recognition image into a pre-trained text recognition model to obtain text information in the text region;
And carrying out structuring processing on the text information in the text region to obtain a text field with a mapping relation.
Further, the structuring processing is performed on the text information in the text area to obtain a text field with a mapping relationship, which specifically includes:
Selecting a preset field from the text information in the text region as a key field, and acquiring the position information of the text region corresponding to the key field;
determining a fuzzy area with a mapping relation with the key field according to the position information of the text area corresponding to the key field;
And detecting and inquiring the text information identified in the fuzzy area, and confirming the text information with the mapping relation with the key field.
Further, the determining, according to the location information of the text region corresponding to the key field, the fuzzy region having a mapping relationship with the key field specifically includes:
Moving the text region corresponding to the key field by a preset distance along the horizontal and vertical directions, and acquiring the position information of the text region after moving according to the position information of the text region corresponding to the key field;
And amplifying the text region after the preset distance is moved based on the position information of the text region after the movement, and determining a fuzzy region with a mapping relation with the key field.
Further, the detecting and querying the text information identified in the fuzzy area, and confirming the text information having a mapping relation with the key field specifically includes:
Detecting position information of all text areas in the fuzzy area, and extracting text information of all text areas in the fuzzy area;
traversing the text information of each text region in the fuzzy region in a regular matching mode, and confirming the text information with a mapping relation with the key field.
Further, traversing the text information of each text region in the fuzzy region in a regular matching manner, and confirming the text information with a mapping relation with the key field, wherein the method specifically comprises the following steps:
Constructing a regular expression matched with the key field by acquiring pattern characters applicable to the key field;
And verifying the text information of each text region in the fuzzy region according to the regular expression matched with the key field, and confirming the text information with the mapping relation with the key field.
Further, before the determining, based on the pre-trained text region detection model, the location information of the text region in the target recognition image, the method further includes:
The collected image sample data is input into a network model for training after text region labeling, and a text region detection model is obtained;
The network model comprises a multi-layer structure, the collected image sample data is input into the network model for training after text region labeling, and a text region detection model is obtained, and the method specifically comprises the following steps:
Extracting image region features corresponding to image sample data through a convolution layer of the network model;
Generating horizontal text sequence features according to the image region features corresponding to the image sample data through a decoding layer of the network model;
and determining a text region in the image sample data according to the horizontal text sequence characteristics by a prediction layer of the network model, and processing the text region to obtain a candidate text line.
Further, the prediction layer of the network model includes a classification part and a regression part, and the method for determining a text region in the image sample data according to the horizontal text sequence feature by the prediction layer of the network model, and processing the text region to obtain a candidate text line specifically includes:
Classifying each region in the image sample data according to the horizontal text sequence features by a classification part of a prediction layer of the network model, and determining a text region in the image sample data;
And carrying out frame regression processing on the text region in the image text data through the regression part of the prediction layer of the network model to obtain candidate text lines.
According to another aspect of the present invention, there is provided an image text recognition apparatus, the apparatus comprising:
The acquisition unit is used for acquiring an image to be identified, and preprocessing the image to be identified to obtain a target identification image;
a determining unit, configured to determine location information of a text region in the target recognition image based on a text region detection model trained in advance;
The recognition unit is used for inputting the target recognition image and the position information of the text area in the target recognition image into a pre-trained text recognition model to obtain text information in the text area;
And the processing unit is used for carrying out structuring processing on the text information in the text region to obtain a text field with a mapping relation.
Further, the processing unit includes:
The selecting module is used for selecting a preset field from the text information in the text area as a key field and acquiring the position information of the text area corresponding to the key field;
the determining module is used for determining a fuzzy area with a mapping relation with the key field according to the position information of the text area corresponding to the key field;
And the detection module is used for detecting and inquiring the text information identified in the fuzzy area and confirming the text information with the mapping relation with the key field.
Further, the determining module includes:
The obtaining sub-module is used for moving the text area corresponding to the key field by a preset distance along the horizontal and vertical directions, and obtaining the position information of the text area after the movement according to the position information of the text area corresponding to the key field;
and the determining submodule is used for amplifying the text area after the preset distance is moved based on the position information of the text area after the movement and determining a fuzzy area with a mapping relation with the key field.
Further, the detection module includes:
the extraction submodule is used for detecting the position information of all text areas in the fuzzy area and extracting the text information of all text areas in the fuzzy area;
and the confirming sub-module is used for traversing the text information of each text area in the fuzzy area in a regular matching mode and confirming the text information with the mapping relation with the key field.
Further, the confirmation sub-module is specifically configured to construct a regular expression matched with the key field by acquiring a pattern character applicable to the key field;
and the confirmation sub-module is specifically further used for verifying the text information of each text area in the fuzzy area according to the regular expression matched with the key field, and confirming the text information with the mapping relation with the key field.
Further, the apparatus further comprises:
The training unit is used for marking the text region of the collected image sample data and inputting the marked text region into the network model for training before the text region detection model based on the pre-training is determined and the position information of the text region in the target identification image is determined, so that the text region detection model is obtained;
The network model comprises a multi-layer structure, and the training unit comprises:
The extraction module is used for extracting image region features corresponding to the image sample data through the convolution layer of the network model;
the generation module is used for generating horizontal text sequence features according to the image region features corresponding to the image sample data through a decoding layer of the network model;
And the prediction module is used for determining a text region in the image sample data according to the horizontal text sequence characteristics through a prediction layer of the network model, and processing the text region to obtain a candidate text line.
Further, the prediction layer of the network model includes a classification portion and a regression portion, and the prediction module includes:
A classification sub-module, configured to classify each region in the image sample data according to the horizontal text sequence feature by using a classification part of a prediction layer of the network model, and determine a text region in the image sample data;
And the processing submodule is used for carrying out frame regression processing on the text region in the image text data through the regression part of the prediction layer of the network model to obtain candidate text lines.
According to a further aspect of the present invention there is provided a computer device comprising a memory storing a computer program and a processor implementing the steps of the image text recognition method when the computer program is executed by the processor.
According to a further aspect of the present invention there is provided a computer storage medium having stored thereon a computer program which when executed by a processor performs the steps of a method of image text recognition.
By means of the technical scheme, the image text recognition method and device provided by the application are used for obtaining the image to be recognized, preprocessing the image to be recognized to obtain the target recognition image, determining the position information of the text region in the target recognition image based on the pre-trained text region detection model, inputting the position information of the text region in the target recognition image and the target recognition image into the pre-trained text recognition model, recognizing the text information in the text region, and structuring the text information of the text region into text fields with mapping relations. Compared with the image text recognition method in the prior art, the method can effectively remove the interference information in the image and accurately reserve the text information of the image by carrying out the structuring processing on the text information after recognition, so that the image detected in the text area and recognized by the text information is not interfered by the background, the corresponding relation of different fields is output and realized, and the accuracy of the image text recognition is improved.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to designate like parts throughout the figures. In the drawings:
Fig. 1 shows a schematic flow chart of an image text recognition method according to an embodiment of the present invention;
Fig. 2 is a schematic flow chart of another image text recognition method according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an image text recognition device according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of another image text recognition device according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
The embodiment of the invention provides an image text recognition method, which can carry out security authentication on a service calling party and ensure the security of service calling, and as shown in figure 1, the method comprises the following steps:
101. And acquiring an image to be identified, and preprocessing the image to be identified to obtain a target identification image.
The image to be identified can be an invoice image, an advertisement image, a commodity image and the like. The preprocessing of the image to be identified herein may include, but is not limited to, rotation correction of the image, removal of background interference from the image.
It will be appreciated that the user may not consider the direction of the image and the shooting angle when uploading the image to be identified, for example, some users are used to use landscape shooting and some users are used to use portrait shooting. In order to facilitate the subsequent processing of the images, the images need to be subjected to angle normalization, a 4-classification model can be trained in advance by using a Resnet network based on a neural network, and the classification model can give prediction angles of 0 degrees, 90 degrees, 180 degrees and 270 degrees after the images to be recognized are input, and the images to be recognized are subjected to rotation correction according to the prediction angles output by the classification model.
The training process of the specific classification model can be realized by firstly preparing 4 types of images which are respectively 0 degree, 90 degrees, 180 degrees and 270 degrees and corresponding angle labels as training data, inputting the training data into a Resnet network, extracting the characteristics of each image by the network, predicting the corresponding angles, carrying out back propagation based on the deviation between the predicted value and the actual value of the angle labels as loss, updating parameters in the network, retaining forward propagation after the training of the classification model is completed, carrying out angle prediction on the images to be identified to obtain the predicted angles, and rotating and correcting all the images to be identified which are not 0 degrees to be 0 degrees to ensure that the predicted angles corresponding to the images to be identified are unified to be 0 degrees, thereby obtaining the images to be identified with the same angle.
It can be appreciated that the image to be identified uploaded by the user generally contains complex background information, some background textures are very similar to text textures, some even contain text, the backgrounds cause great interference to identification and detection, the background interference of the image needs to be removed, the background and the foreground of the image can be segmented by combining the image segmentation technology based on a network model built by DeepLab-v3, the minimum external rectangle of the foreground is calculated, and the background is cut off along the minimum external matrix.
The implementation process of image background and foreground segmentation of the image to be identified based on DeepLab-v3 constructed network model combined with image segmentation technology can be as follows, firstly, training data are prepared, an image dataset and corresponding labels, the corresponding pixel size of each image label is consistent with the original image size, the pixel value corresponding to the label is 0 to represent the image background, the pixel value corresponding to the label is 255 to represent the image foreground, deepLab-v3 network is built, training data are input, classification prediction is carried out on each pixel, loss between a predicted value and a real pixel value of the label is calculated, then the parameters in the network are reversely propagated, and the parameters in the network are updated until the accuracy evaluation index mean-IOU of the network model reaches a preset value. And identifying pixels in the image to be identified based on the constructed network model to obtain marks of foreground and background in the image, namely identifying the mark with the pixel value of 0 as the image background, identifying the foreground with the pixel value of 255 as the image, segmenting the image to be identified based on the mark of the background, and removing the background of the image to be identified.
The image to be identified after pretreatment can better express the image characteristics, so that the image to be identified is used as a target image to further identify the text in the image.
102. And determining the position information of the text region in the target identification image based on a pre-trained text region detection model.
The pre-trained text box detection model may use an open source DETECTING TEXT IN NATURAL IMAGE WITH Connectionist Text Proposal Network (CTPN) framework, and each target identification image has a corresponding output file after passing through the text region detection model, where the output file stores position information obtained by detecting all text regions, and the position information of the text regions in the target identification image can be determined through the output file.
The process of training the text region detection model can be that training data, namely an image and a label file corresponding to the image are prepared firstly, coordinate information of a text region in the image is stored in the label file, in order to facilitate subsequent detection of the text region in the image, before each training data is input to a CTPN network, the coordinate information marked by the text region is required to be converted into a small anchor with the width of 8, CTPN network structure adopts a form of CNN+BLSTM+RPN, CNN is used for extracting spatial characteristics of a receptive field in the image, the receptive field is a region of an input image corresponding to a certain node which is convolved by a convolution kernel, BLSTM can generate horizontal text sequence characteristics based on the spatial characteristics of the receptive field, the anchor classification and bounding box regressing regression are carried out, and after bounding box regressing regression processing, a group of vertical strip candidate text lines can be obtained after the anchor classification can be used for determining whether each region is the text region or not.
It should be noted that, the text region detection model after training is not directly output to the text region in the target recognition image, but a group of vertical strip candidate text lines forming the text region in the target recognition image, and a text line construction algorithm may be used to connect the group of vertical strip candidate text lines into the text region, so as to determine the text region in the target recognition image and the position information of the text region.
103. And inputting the target recognition image and the position information of the text region in the target recognition image into a pre-trained text recognition model to obtain the text information in the text region.
The text recognition model can train the recognition model by adopting An End-to-End Trainable Neural Network for Imaged-based Sequence Recognition and Its Application to Scene Text R ecognition(CRNN) algorithm, and after the position information of the text region in the target recognition image and the target recognition image passes through the text recognition model, the text recognition result corresponding to each text region is output.
The specific process of training the CRNN model can be that firstly training data is stored in a label mode of images and text information of text areas in the images, a CRNN network structure is in a form of CNN+RNN+CTC, the CNN is used for extracting spatial characteristics of receptive fields in the images, the RNN can predict label distribution of each frame in the images based on the spatial characteristics of the receptive fields, and the CTC can integrate the label distribution of each frame into a final label sequence. For example, the size of the input picture resize to w×32, and the predicted value output through the text recognition model represents text information corresponding to the text region in the target recognition image.
104. And carrying out structuring processing on the text information in the text region to obtain a text field with a mapping relation.
There may be various types, for example, included in the text information in the text region due to the recognition model. Text type, number type, special character type, etc. and there may be a mapping relationship between them, for example, a "37.5-membered whole" text region corresponding to a "payment amount" text region, a "female" text region corresponding to a "gender" text box, and by performing structured output on text information, the mapping relationship between text information can be clearly seen.
Specifically, text information identified by a specific text region in an image can be used as a key field, a fuzzy region corresponding to the text field with a mapping relation can be located based on the key field, then the field with the mapping relation with the key field is locked in a region range, and further the text information in the region range is checked, so that the text information with the mapping relation with the key field is confirmed, and similarly, other fields in the image can be used in the same way, and further the text field with the mapping relation can be obtained.
According to the image text recognition method provided by the embodiment of the application, the image to be recognized is obtained, the image to be recognized is preprocessed to obtain the target recognition image, the position information of the text region in the target recognition image is determined based on the pre-trained text region detection model, after the position information of the text region in the target recognition image and the target recognition image is input into the pre-trained text recognition model, the text information in the text region is obtained through recognition, and the text information of the text region is structured into the text field with the mapping relation. Compared with the image text recognition method in the prior art, the method can effectively remove the interference information in the image and accurately reserve the text information of the image by carrying out the structuring processing on the text information after recognition, so that the image detected in the text area and recognized by the text information is not interfered by the background, the corresponding relation of different fields is output and realized, and the accuracy of the image text recognition is improved.
The embodiment of the invention provides another image text recognition method, which can carry out security authentication on a service calling party and ensure the security of service calling, and as shown in fig. 2, the method comprises the following steps:
201. and acquiring an image to be identified, and preprocessing the image to be identified to obtain a target identification image.
For the embodiment of the present invention, specifically, the image to be identified is obtained, the image to be identified is preprocessed, and the process of obtaining the target identification image may refer to the content described in step 101, which is not described herein.
202. And (3) marking the text region of the collected image sample data, and inputting the text region marked image data into a network model for training to obtain a text region detection model.
The image sample data may be images collected from different scenes, and may reflect image features in different scenes, for example, the image features in the medical field are simpler or gradually changed, the image background scene in the industrial field is more complex, the text area is relatively smaller, the image background of the natural scene is affected by natural factors more, and the background complexity is difficult to predict.
It will be appreciated that in any of the scenarios of the image, in the general object detection, each object has a clear closed boundary, and in the image, since the text line or word is composed of a plurality of individual characters or strokes, there may not be such a clear defined boundary, it is necessary to detect the text region in the image first, specifically, by labeling the text region contained in each image in the image sample data, training the labeled image sample data, constructing a text region detection model, and detecting the text region in the image by using the text region detection model, thereby identifying the text in the image.
For the embodiment of the invention, a network model can adopt a CTPN network frame, and comprises a 3-layer structure, wherein the first layer is in a convolution structure, namely a CNN structure, spatial information of a receptive field can be learned by extracting image region characteristics corresponding to image sample data through the convolution layer, the second layer is in a decoding layer, namely a BLSTM structure, horizontal text sequence characteristics can be well processed by generating the horizontal text sequence characteristics through the decoding layer according to the image region characteristics corresponding to the image sample data, and the third layer is in a prediction layer, namely an RPN structure, text regions in the image sample data are determined through the prediction layer according to the horizontal text sequence characteristics, and the text regions are processed to obtain candidate text lines.
Specifically, the prediction layer of the network model comprises a classification part and a regression part, and in the process of determining text areas in image sample data according to horizontal text sequence characteristics through the prediction layer of the network model and processing the text areas to obtain candidate text lines, the classification part of the prediction layer of the network model can be used for classifying each area in the image sample data according to the horizontal text sequence characteristics to determine the text areas in the image sample data, and the regression part of the prediction layer of the network model is used for carrying out frame regression processing on the text areas in the image text data to obtain the candidate text lines.
In the implementation process, feature maps of conv5 in the VGG model can be selected as the final feature of the image in the part CTPN of the convolution layer, the size of feature maps is H×W×C, then, due to the sequence relation among texts, 3×3 sliding windows can be adopted in the decoding layer to extract 3×3 areas around each point on feature maps as the feature vector representation of the point, at the moment, the size of the image is changed into H×W×9C, then, each row is used as the length of the sequence, the height is used as the batch_size, a 128-dimensional Bi-LSTM is transmitted, the output of the decoding layer is W×H×256, finally, the decoding layer is output and connected to a prediction layer, the prediction layer comprises two parts, the anchor classification and bounding box regressing, whether each area in the image is a text area can be determined through the anchor classification, a group of vertical strip candidate text lines can be obtained after bounding box regressing processing, and whether the candidate text lines are labels of the text areas or not can be carried.
Further, in order to ensure the accuracy of prediction of the text region detection model obtained through training, the preset loss function can carry out parameter adjustment on the multi-layer structure in the text region detection model based on the deviation between the result output by the text region detection model and the data marked by the real text region. For the embodiment of the invention, the pre-trained loss function comprises 3 parts, wherein the first part is a loss function for detecting whether the Anchor is a text region, the second part is a loss function for detecting y-coordinate offset regression of the Anchor, and the third part is a loss function for detecting x-coordinate offset regression of the Anchor.
203. And determining the position information of the text region in the target identification image based on a pre-trained text region detection model.
It can be understood that each image has a corresponding output file through the text region detection model, the output file stores the position information of all candidate text lines in the image and whether the candidate text lines are labels of the text regions, the candidate text lines correspond to vertical strip lines split from the text regions, and the position information of the text regions in the target image can be determined by looking at the candidate text line connection.
204. And inputting the target recognition image and the position information of the text region in the target recognition image into a pre-trained text recognition model to obtain the text information in the text region.
It can be appreciated that the trained text recognition model has the capability of recognizing text information in a text region, and because the sample image and the position information label of the text region in the sample image are used in the process of training the text recognition model, parameters of the text recognition model are continuously adjusted through forward propagation and reverse deviation correction, so that the text information in the text region can be accurately recognized through the image of the text recognition model.
205. And selecting a preset field from the text information in the text area as a key field, and acquiring the position information of the text area corresponding to the key field.
As the key field is usually a field with reference value in the image, for the image of the invoice, the fields such as the invoice number, the total amount, the medical insurance type, the number, the date and time and the like can be selected as preset fields and used as the key field, and the coordinate information corresponding to the key field can be obtained through a text region detection model.
206. And determining a fuzzy area with a mapping relation with the key field according to the position information of the text area corresponding to the key field.
For the embodiment of the present invention, in order to accurately locate text information mapped to a key field, since text information having a mapping relationship with the key field is usually located at one side of the key field, the text information mapped to the key field may be located in a range within a fuzzy area by determining a fuzzy area having a mapping relationship with the key field, for example, coordinate information of a text area corresponding to the key field may be [ (Xmin, ymin), (Xmax, ymax) ], where (Xmin, ymin) is a coordinate of an upper left corner and (Xmax, ymax) is a coordinate of a lower right corner, then coordinate information of a fuzzy area having a mapping relationship with the key field may be [ (Xmin, ymin+2/3 (Ymax-Ymin)), (xmin+1/2 (Xmax-Xmin)), ymax) ], where (xmin+2/3 (Ymax-Ymin)) is a coordinate of an upper left corner and (xmin+1/2 (Xmax-Xmin)) is a coordinate of a lower right corner.
Specifically, the text region corresponding to the key field is moved by a preset distance along the horizontal direction and the vertical direction, the position information of the text region after the movement is acquired according to the position information of the text region corresponding to the key field, for example, the text region after the movement is moved by a distance of 1/2 (Xmax-Xmin) along the horizontal direction and a distance of 2/3 (Ymax-Ymin) along the vertical direction, and because the text region after the movement is possibly not covered by the text information with the mapping relation with the key field, the text region after the movement is amplified based on the position information of the text region after the movement, and the fuzzy region with the mapping relation with the key field is determined.
207. And detecting and inquiring the text information identified in the fuzzy area, and confirming the text information with the mapping relation with the key field.
Specifically, the text information of all text areas in the fuzzy area is detected and extracted, and the text information of each text area in the fuzzy area is further traversed in a regular matching mode to confirm the text information with the mapping relation with the key field because the text information of all text areas in the fuzzy area is not text information with the mapping relation with the key field.
For the embodiment of the invention, in the process of traversing the text information of each text region in the fuzzy region in a regular matching mode and confirming the text information with the mapping relation with the key field, the method can be particularly used for judging whether the text information has the mapping relation with the key field or not by acquiring the pattern character suitable for the key field, constructing a regular expression matching the key field, for example, for the case where the aggregate amount is the key field, the pattern characters applicable to the aggregate amount may include zero, one, two, three, four, five, land, seven, eight, jiu, ten, bai, yuan, round, etc., the structured regular expression can be but is not "([ Whole zero one three four Wu Liu eight red 1,2 ten) 1, 1} ([ whole zero one three four Wu Liu eight red 1,2} Qia Jiu ] {1,2} Qia) {0,1} ([ whole zero one three four Wu Liu eight red 1,2} Bai) {0,1} ([ whole zero one three red 1,2} Bai) {0,1 }; four Wu Liu eight-Jiu ] {1,2} pickup) {0,1} zero {0,1} ([ whole zero one two three four Wu Liu eight-eight ] {0,2} [ Yuanyuan ]) {0,1} ([ whole zero one three four Wu Liu eight-eight } [ Yuanyuan ]) } ([ zero three four eight-four eight-four-eight-eight one two three four Wu Liu eight (1, 2) angle) {0,1} ([ Whole zero one two three four Wu Liu eight-way ] {1,2} branch) {0,1} whole {0,1 }). And verifying the text information of each text region in the fuzzy region according to the regular expression matched with the key field, and confirming the text information with the mapping relation with the key field, so that the text information meeting the regular expression is the text information with the mapping relation with the key field.
It should be noted that, because the field formats and types in different text information are different, for some fields, the structural confirmation is performed in a manner that the field is suitable for regular matching, the structural confirmation of other fields may be performed in a manner of coordinate calculation, specifically, the structural confirmation may be performed by calculating, based on the lower left corner coordinates of the key field, a text region nearest to the point as a text region of the next key field by taking the height of the text region of the key field as a calculation unit of point movement, and further determining, based on the text region of the next key field, a text field having a mapping relationship with the key field.
Further, as a specific implementation of the method shown in fig. 1, an embodiment of the present invention provides an image text recognition device, as shown in fig. 3, where the device includes an obtaining unit 31, a determining unit 32, a recognition unit 33, and a processing unit 34.
The acquiring unit 31 may be configured to acquire an image to be identified, and perform preprocessing on the image to be identified to obtain a target identification image;
A determining unit 32, configured to determine location information of a text region in the target recognition image based on a text region detection model trained in advance;
the recognition unit 33 may be configured to input the target recognition image and the position information of the text region in the target recognition image into a pre-trained text recognition model, to obtain text information in the text region;
the processing unit 34 may be configured to perform a structuring process on the text information in the text area, so as to obtain a text field with a mapping relationship.
According to the image text recognition device provided by the embodiment of the application, the image to be recognized is obtained, the image to be recognized is preprocessed to obtain the target recognition image, the position information of the text region in the target recognition image is determined based on the pre-trained text region detection model, after the position information of the text region in the target recognition image and the target recognition image is input into the pre-trained text recognition model, the text information in the text region is obtained through recognition, and the text information of the text region is structured into the text field with the mapping relation. Compared with the image text recognition method in the prior art, the method can effectively remove the interference information in the image and accurately reserve the text information of the image by carrying out the structuring processing on the text information after recognition, so that the image detected in the text area and recognized by the text information is not interfered by the background, the corresponding relation of different fields is output and realized, and the accuracy of the image text recognition is improved.
As a further explanation of the image text recognition apparatus shown in fig. 3, fig. 4 is a schematic structural view of another image text recognition apparatus according to an embodiment of the present invention, and as shown in fig. 4, the processing unit 34 includes:
The selecting module 341 may be configured to select a preset field from the text information in the text area as a key field, and obtain location information of a text area corresponding to the key field;
The determining module 342 may be configured to determine, according to the location information of the text region corresponding to the key field, a fuzzy region having a mapping relationship with the key field;
The detection module 343 may be configured to detect and query the text information identified in the fuzzy area, and confirm the text information having a mapping relationship with the key field.
Further, the determining module 342 includes:
The acquiring submodule 3421 is configured to move the text region corresponding to the key field by a preset distance along the horizontal and vertical directions, and acquire the position information of the moved text region according to the position information of the text region corresponding to the key field;
The determining submodule 3422 may be configured to perform amplification processing on the text region after the movement by a preset distance based on the position information of the text region after the movement, and determine a fuzzy region having a mapping relationship with the key field.
Further, the detection module 343 includes:
An extraction submodule 3431 which can be used for detecting the position information of all text areas in the fuzzy area and extracting the text information of all text areas in the fuzzy area;
The confirmation sub-module 3432 may be configured to traverse the text information of each text region in the fuzzy region in a regular matching manner, and confirm the text information having a mapping relationship with the key field.
Further, the confirmation submodule 3432 may be specifically configured to construct a regular expression matched with the key field by acquiring a pattern character applicable to the key field;
The confirmation submodule 3432 may be further configured to verify the text information of each text region in the fuzzy region according to the regular expression matched with the key field, and confirm the text information having a mapping relationship with the key field.
Further, the apparatus further comprises:
The training unit 35 may be configured to, before determining the location information of the text region in the target recognition image based on the pre-trained text region detection model, perform text region labeling on the collected image sample data, and then input the labeled text region sample data into a network model for training, so as to obtain a text region detection model;
the network model includes a multi-layer structure, and the training unit 35 includes:
The extracting module 351 may be configured to extract, by using a convolution layer of the network model, an image region feature corresponding to image sample data;
the generating module 352 may be configured to generate, by using a decoding layer of the network model, a horizontal text sequence feature according to an image region feature corresponding to the image sample data;
The prediction module 353 may be configured to determine, by using a prediction layer of the network model, a text region in the image sample data according to the horizontal text sequence feature, and process the text region to obtain a candidate text line.
Further, the prediction layer of the network model includes a classification part and a regression part, and the prediction module 353 includes:
A classification sub-module 3531, configured to classify, by a classification portion of a prediction layer of the network model, respective regions in the image sample data according to the horizontal text sequence feature, and determine text regions in the image sample data;
And a processing submodule 3532, configured to perform frame regression processing on a text region in the image text data through a regression portion of a prediction layer of the network model, so as to obtain a candidate text line.
It should be noted that, other corresponding descriptions of each functional unit related to the image text recognition device provided in this embodiment may refer to corresponding descriptions in fig. 1 and fig. 2, and are not described herein again.
Based on the above-mentioned methods shown in fig. 1 and 2, correspondingly, the present embodiment further provides a storage medium, on which a computer program is stored, which when executed by a processor, implements the above-mentioned image text recognition method shown in fig. 1 and 2.
Based on such understanding, the technical solution of the present application may be embodied in the form of a software product, which may be stored in a non-volatile storage medium (may be a CD-ROM, a U-disk, a mobile hard disk, etc.), and includes several instructions for causing a computer device (may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective implementation scenario of the present application.
Based on the methods shown in fig. 1 and fig. 2 and the virtual device embodiments shown in fig. 3 and fig. 4, in order to achieve the above objects, the embodiments of the present application further provide a computer device, which may specifically be a personal computer, a server, a network device, or the like, where the entity device includes a storage medium and a processor, where the storage medium is used to store a computer program, and where the processor is used to execute the computer program to implement the image text recognition method shown in fig. 1 and fig. 2.
Optionally, the computer device may also include a user interface, a network interface, a camera, radio Frequency (RF) circuitry, sensors, audio circuitry, WI-FI modules, and the like. The user interface may include a Display screen (Display), an input unit such as a Keyboard (Keyboard), etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (e.g., bluetooth interface, WI-FI interface), etc.
It will be appreciated by those skilled in the art that the structure of the entity device of the image text recognition apparatus provided in this embodiment is not limited to the entity device, and may include more or fewer components, or may combine some components, or may be different in arrangement of components.
The storage medium may also include an operating system, a network communication module. An operating system is a program that manages the computer device hardware and software resources described above, supporting the execution of information handling programs and other software and/or programs. The network communication module is used for realizing communication among all components in the storage medium and communication with other hardware and software in the entity equipment.
From the above description of the embodiments, it will be apparent to those skilled in the art that the present application may be implemented by means of software plus necessary general hardware platforms, or may be implemented by hardware. By applying the technical scheme of the application, compared with the prior art, the method and the device can effectively remove the interference information in the image and accurately reserve the text information of the image by carrying out the structuring treatment on the text information after the recognition, so that the image detected in the text area and recognized by the text information is not interfered by the background, thereby outputting the corresponding relation of different fields and improving the accuracy of the text recognition of the image.
Those skilled in the art will appreciate that the drawing is merely a schematic illustration of a preferred implementation scenario and that the modules or flows in the drawing are not necessarily required to practice the application. Those skilled in the art will appreciate that modules in an apparatus in an implementation scenario may be distributed in an apparatus in an implementation scenario according to an implementation scenario description, or that corresponding changes may be located in one or more apparatuses different from the implementation scenario. The modules of the implementation scenario may be combined into one module, or may be further split into a plurality of sub-modules.
The above-mentioned inventive sequence numbers are merely for description and do not represent advantages or disadvantages of the implementation scenario. The foregoing disclosure is merely illustrative of some embodiments of the application, and the application is not limited thereto, as modifications may be made by those skilled in the art without departing from the scope of the application.
Claims (7)
1. A method of image text recognition, the method comprising:
Acquiring an image to be identified, and preprocessing the image to be identified to obtain a target identification image;
Determining the position information of a text region in the target recognition image based on a pre-trained text region detection model;
Inputting the target recognition image and the position information of the text region in the target recognition image into a pre-trained text recognition model to obtain text information in the text region;
The method comprises the steps of carrying out structuring processing on text information in a text area to obtain text fields with mapping relation, specifically, selecting preset fields from the text information in the text area to serve as key fields, obtaining the position information of the text area corresponding to the key fields, moving the text area corresponding to the key fields along the horizontal and vertical directions by preset distances, obtaining the position information of the moved text area according to the position information of the text area corresponding to the key fields, carrying out amplifying processing on the text area after the preset distance is moved based on the position information of the text area after the movement, determining fuzzy areas with mapping relation with the key fields, detecting the position information of all the text areas in the fuzzy areas, extracting the text information of all the text areas in the fuzzy areas, traversing the text information of each text area in the fuzzy areas in a regular matching mode, and confirming the text information with mapping relation with the key fields, wherein the preset fields are fields with reference values in images.
2. The method according to claim 1, wherein traversing the text information of each text region in the fuzzy region by using a regular matching method, and confirming the text information having a mapping relation with the key field specifically comprises:
Constructing a regular expression matched with the key field by acquiring pattern characters applicable to the key field;
And verifying the text information of each text region in the fuzzy region according to the regular expression matched with the key field, and confirming the text information with the mapping relation with the key field.
3. The method according to any one of claims 1-2, wherein prior to the determining location information of text regions in the target recognition image based on a pre-trained text region detection model, the method further comprises:
The collected image sample data is input into a network model for training after text region labeling, and a text region detection model is obtained;
The network model comprises a multi-layer structure, the collected image sample data is input into the network model for training after text region labeling, and a text region detection model is obtained, and the method specifically comprises the following steps:
Extracting image region features corresponding to image sample data through a convolution layer of the network model;
Generating horizontal text sequence features according to the image region features corresponding to the image sample data through a decoding layer of the network model;
and determining a text region in the image sample data according to the horizontal text sequence characteristics by a prediction layer of the network model, and processing the text region to obtain a candidate text line.
4. A method according to claim 3, wherein the prediction layer of the network model comprises a classification part and a regression part, the prediction layer of the network model determines text regions in the image sample data according to the horizontal text sequence features, and processes the text regions to obtain candidate text lines, and the method specifically comprises:
Classifying each region in the image sample data according to the horizontal text sequence features by a classification part of a prediction layer of the network model, and determining a text region in the image sample data;
And carrying out frame regression processing on the text region in the image text data through the regression part of the prediction layer of the network model to obtain candidate text lines.
5. An image text recognition device, the device comprising:
The acquisition unit is used for acquiring an image to be identified, and preprocessing the image to be identified to obtain a target identification image;
a determining unit, configured to determine location information of a text region in the target recognition image based on a text region detection model trained in advance;
The recognition unit is used for inputting the target recognition image and the position information of the text area in the target recognition image into a pre-trained text recognition model to obtain text information in the text area;
The processing unit is used for carrying out structural processing on the text information in the text region to obtain text fields with mapping relation, and concretely comprises the steps of selecting preset fields from the text information in the text region as key fields, obtaining the position information of the text region corresponding to the key fields, moving the text region corresponding to the key fields along the horizontal and vertical directions by preset distances, obtaining the position information of the text region after moving according to the position information of the text region corresponding to the key fields, carrying out amplification processing on the text region after moving by preset distances based on the position information of the text region after moving, determining fuzzy regions with mapping relation with the key fields, detecting the position information of all the text regions in the fuzzy regions, extracting the text information of all the text regions in the fuzzy regions, traversing the text information of each text region in the fuzzy regions in a regular matching mode, and confirming the text information with mapping relation with the key fields, wherein the preset fields are fields with reference values in images.
6. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 4 when the computer program is executed.
7. A computer storage medium having stored thereon a computer program, which when executed by a processor realizes the steps of the method according to any of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010051370.3A CN111259889B (en) | 2020-01-17 | 2020-01-17 | Image text recognition method, device, computer equipment and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010051370.3A CN111259889B (en) | 2020-01-17 | 2020-01-17 | Image text recognition method, device, computer equipment and computer storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111259889A CN111259889A (en) | 2020-06-09 |
| CN111259889B true CN111259889B (en) | 2025-03-14 |
Family
ID=70923742
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010051370.3A Active CN111259889B (en) | 2020-01-17 | 2020-01-17 | Image text recognition method, device, computer equipment and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111259889B (en) |
Families Citing this family (33)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112016561B (en) * | 2020-09-01 | 2023-08-04 | 中国银行股份有限公司 | Text recognition method and related equipment |
| CN111798542B (en) * | 2020-09-10 | 2020-12-22 | 北京易真学思教育科技有限公司 | Model training method, data processing device, model training apparatus, and storage medium |
| CN112149654B (en) * | 2020-09-23 | 2022-08-02 | 四川长虹电器股份有限公司 | Invoice text information identification method based on deep learning |
| CN112052835B (en) | 2020-09-29 | 2022-10-11 | 北京百度网讯科技有限公司 | Information processing method, information processing apparatus, electronic device, and storage medium |
| CN114549818A (en) * | 2020-11-25 | 2022-05-27 | 上海云扩信息科技有限公司 | A method and device for extracting structured information from super large images |
| CN112396054A (en) * | 2020-11-30 | 2021-02-23 | 泰康保险集团股份有限公司 | Text extraction method and device, electronic equipment and storage medium |
| CN112651392A (en) * | 2020-12-22 | 2021-04-13 | 深圳壹账通智能科技有限公司 | Certificate information acquisition method and device, storage medium and computer equipment |
| CN112668583A (en) * | 2021-01-07 | 2021-04-16 | 浙江星汉信息技术股份有限公司 | Image recognition method and device and electronic equipment |
| CN112818645A (en) * | 2021-02-02 | 2021-05-18 | 广州楹鼎生物科技有限公司 | Chemical information extraction method, device, equipment and storage medium |
| CN113569082B (en) * | 2021-02-08 | 2025-05-27 | 腾讯科技（深圳）有限公司 | Image information output method and related device |
| CN112926471B (en) * | 2021-03-05 | 2025-03-21 | 中国工商银行股份有限公司 | Business document image content recognition method and device |
| CN112633283A (en) * | 2021-03-08 | 2021-04-09 | 广州市玄武无线科技股份有限公司 | Method and system for identifying and translating English mail address |
| CN113012189B (en) * | 2021-03-31 | 2024-07-26 | 影石创新科技股份有限公司 | Image recognition method, device, computer equipment and storage medium |
| CN113128490B (en) * | 2021-04-28 | 2023-12-05 | 湖南荣冠智能科技有限公司 | Prescription information scanning and automatic identification method |
| CN115512375A (en) * | 2021-06-07 | 2022-12-23 | 中移物联网有限公司 | Text error correction model training method, text recognition method and related equipment |
| CN113408536A (en) * | 2021-06-23 | 2021-09-17 | 平安健康保险股份有限公司 | Bill amount identification method and device, computer equipment and storage medium |
| CN113591657B (en) * | 2021-07-23 | 2024-04-09 | 京东科技控股股份有限公司 | OCR layout recognition method and device, electronic equipment and medium |
| CN113705449B (en) * | 2021-08-27 | 2025-01-17 | 上海商汤临港智能科技有限公司 | Identification recognition method and related device |
| CN113869316B (en) * | 2021-09-10 | 2025-03-25 | 深圳前海微众银行股份有限公司 | Information notification method, device, equipment and computer storage medium |
| CN113887375A (en) * | 2021-09-27 | 2022-01-04 | 浙江大华技术股份有限公司 | Text recognition method, device, equipment and storage medium |
| CN116109543A (en) * | 2021-11-10 | 2023-05-12 | 中移(成都)信息通信科技有限公司 | Method and device for quickly identifying and reading data and computer readable storage medium |
| CN114038004A (en) * | 2021-11-19 | 2022-02-11 | 北京恒安嘉新安全技术有限公司 | A method, device, device and storage medium for extracting certificate information |
| CN114329030A (en) * | 2021-11-25 | 2022-04-12 | 腾讯科技（深圳）有限公司 | Information processing method and device, computer equipment and storage medium |
| CN114463767B (en) * | 2021-12-28 | 2024-11-26 | 上海浦东发展银行股份有限公司 | Letter of credit identification method, device, computer equipment and storage medium |
| CN114358199A (en) * | 2022-01-10 | 2022-04-15 | 上海亿保健康管理有限公司 | Training method of lightweight character recognition model, electronic device and storage medium |
| CN114548096A (en) * | 2022-01-20 | 2022-05-27 | 钻技(上海)信息科技有限公司 | Merchant information acquisition method and system and readable storage medium |
| CN114495126A (en) * | 2022-02-11 | 2022-05-13 | 中国工商银行股份有限公司 | Test method and device based on view object recognition and electronic equipment |
| CN114821568B (en) * | 2022-06-27 | 2022-09-16 | 深圳前海环融联易信息科技服务有限公司 | Menu element extraction method and device, computer equipment and storage medium |
| CN115063084B (en) * | 2022-07-12 | 2025-07-25 | 河南省烟草公司信阳市公司 | Cigarette retail merchant inventory checking method and system |
| CN115376137B (en) * | 2022-08-02 | 2023-09-26 | 北京百度网讯科技有限公司 | Optical character recognition processing and text recognition model training method and device |
| CN115512285A (en) * | 2022-10-14 | 2022-12-23 | 杭州朝厚信息科技有限公司 | A parking space number recognition method, device, electronic equipment and storage medium |
| CN115346205A (en) * | 2022-10-17 | 2022-11-15 | 广州简悦信息科技有限公司 | Page information identification method and device and electronic equipment |
| CN118736610B (en) * | 2024-08-19 | 2025-02-07 | 北京荣大科技股份有限公司 | A manuscript image recognition system and method based on fusion detection |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109308476A (en) * | 2018-09-06 | 2019-02-05 | 邬国锐 | Bill information processing method, system and computer readable storage medium |
| CN109635627A (en) * | 2018-10-23 | 2019-04-16 | 中国平安财产保险股份有限公司 | Pictorial information extracting method, device, computer equipment and storage medium |
| CN109857992A (en) * | 2018-12-29 | 2019-06-07 | 医渡云（北京）技术有限公司 | Medical data structuring analytic method, device, readable medium and electronic equipment |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107798299B (en) * | 2017-10-09 | 2020-02-07 | 平安科技（深圳）有限公司 | Bill information identification method, electronic device and readable storage medium |
| CN107766809B (en) * | 2017-10-09 | 2020-05-19 | 平安科技（深圳）有限公司 | Electronic device, ticket information identification method, and computer-readable storage medium |
| CN109086756B (en) * | 2018-06-15 | 2021-08-03 | 众安信息技术服务有限公司 | A method, device and device for text detection and analysis based on deep neural network |
-
2020
- 2020-01-17 CN CN202010051370.3A patent/CN111259889B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109308476A (en) * | 2018-09-06 | 2019-02-05 | 邬国锐 | Bill information processing method, system and computer readable storage medium |
| CN109635627A (en) * | 2018-10-23 | 2019-04-16 | 中国平安财产保险股份有限公司 | Pictorial information extracting method, device, computer equipment and storage medium |
| CN109857992A (en) * | 2018-12-29 | 2019-06-07 | 医渡云（北京）技术有限公司 | Medical data structuring analytic method, device, readable medium and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111259889A (en) | 2020-06-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111259889B (en) | Image text recognition method, device, computer equipment and computer storage medium | |
| US10607362B2 (en) | Remote determination of containers in geographical region | |
| US10319107B2 (en) | Remote determination of quantity stored in containers in geographical region | |
| CN111291629B (en) | Method, device, computer equipment and computer storage medium for recognizing text in images | |
| CN109117848B (en) | Text line character recognition method, device, medium and electronic equipment | |
| CN111275038A (en) | Image text recognition method and device, computer equipment and computer storage medium | |
| US20190294921A1 (en) | Field identification in an image using artificial intelligence | |
| CN110874618A (en) | OCR template learning method and device based on small sample, electronic equipment and medium | |
| CN113642582B (en) | Meter reading recognition method, device, electronic device and storage medium | |
| CN112101386B (en) | Text detection method, device, computer equipment and storage medium | |
| CN111738252B (en) | Text line detection method, device and computer system in image | |
| CN110991201B (en) | Bar code detection method and related device | |
| Shahin et al. | Machine-based identification system via optical character recognition | |
| CN103955499A (en) | Visual experience enhancing method based on instant calculation and dynamic tracing | |
| EP3553700A2 (en) | Remote determination of containers in geographical region | |
| CN112801960B (en) | Image processing method and device, storage medium and electronic equipment | |
| CN116187717A (en) | Method and system for file storage management | |
| Košcevic et al. | Automatic visual reading of meters using deep learning | |
| CN114529912A (en) | Graphic verification code identification method and device, electronic equipment and readable storage medium | |
| CN113486204A (en) | Picture marking method, device, medium and equipment | |
| Vidhyalakshmi et al. | Text detection in natural images with hybrid stroke feature transform and high performance deep Convnet computing | |
| CN112364687A (en) | Improved Faster R-CNN gas station electrostatic sign identification method and system | |
| CN111291756B (en) | Method and device for detecting text region in image, computer equipment and computer storage medium | |
| Dayananda Kumar et al. | Automated parameter-less optical mark recognition | |
| Khin et al. | License Plate Identification from Myanmar Vehicle Images under Different Environmental Conditions. |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20220520 Address after: 518000 China Aviation Center 2901, No. 1018, Huafu Road, Huahang community, Huaqiang North Street, Futian District, Shenzhen, Guangdong Province Applicant after: Shenzhen Ping An medical and Health Technology Service Co.,Ltd. Address before: Room 12G, Area H, 666 Beijing East Road, Huangpu District, Shanghai 200001 Applicant before: PING AN MEDICAL AND HEALTHCARE MANAGEMENT Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |