CN117094966A - Tongue image identification method and device based on image amplification and computer equipment - Google Patents
Tongue image identification method and device based on image amplification and computer equipment Download PDFInfo
- Publication number
- CN117094966A CN117094966A CN202311051530.4A CN202311051530A CN117094966A CN 117094966 A CN117094966 A CN 117094966A CN 202311051530 A CN202311051530 A CN 202311051530A CN 117094966 A CN117094966 A CN 117094966A
- Authority
- CN
- China
- Prior art keywords
- tongue
- image
- model
- images
- training
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 230000003321 amplification Effects 0.000 title abstract description 6
- 238000003199 nucleic acid amplification method Methods 0.000 title abstract description 6
- 230000011218 segmentation Effects 0.000 claims abstract description 64
- 238000012937 correction Methods 0.000 claims abstract description 54
- 238000009792 diffusion process Methods 0.000 claims abstract description 37
- 238000012545 processing Methods 0.000 claims abstract description 27
- 238000012549 training Methods 0.000 claims description 100
- 230000003416 augmentation Effects 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 17
- 238000002372 labelling Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 3
- 238000013145 classification model Methods 0.000 description 16
- 239000011248 coating agent Substances 0.000 description 7
- 238000000576 coating method Methods 0.000 description 7
- 210000001035 gastrointestinal tract Anatomy 0.000 description 5
- 201000010099 disease Diseases 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 239000003086 colorant Substances 0.000 description 3
- 239000003814 drug Substances 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 208000016644 chronic atrophic gastritis Diseases 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000010749 BS 2869 Class C1 Substances 0.000 description 1
- 239000010750 BS 2869 Class C2 Substances 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 208000010643 digestive system disease Diseases 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 208000028774 intestinal disease Diseases 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000004400 mucous membrane Anatomy 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 210000000952 spleen Anatomy 0.000 description 1
- 210000002784 stomach Anatomy 0.000 description 1
- 210000003699 striated muscle Anatomy 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 210000001835 viscera Anatomy 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0012—Biomedical image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/778—Active pattern-learning, e.g. online learning of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
Abstract
The invention relates to the technical field of image processing, and provides a tongue image recognition method, a tongue image recognition device and computer equipment based on image amplification, wherein the method comprises the following steps: carrying out effective region segmentation processing on an initial tongue image to be processed to obtain a first tongue image and a first tongue outline mask image; generating N first generated tongue images corresponding to the first tongue outline mask images based on a preset image diffusion model, wherein N is greater than or equal to 1; respectively carrying out color correction on the first tongue images based on the N first generation type tongue images, and respectively generating N second tongue images corresponding to the N first generation type tongue images one by one; and taking N second tongue images or a set of N second tongue images and the first tongue images as target image data, and calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image. The tongue image recognition method and device improve the accuracy of tongue image recognition.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to a tongue image recognition method, apparatus and computer device based on image amplification.
Background
The tongue is one of the important constituent organs in the human digestive tract, consists of a plurality of crisscrossed striated muscles, and is externally covered with a special mucous membrane structure. In traditional Chinese medicine, the tongue is closely related to the viscera, especially the spleen and stomach, so that the tongue is often used to infer the digestive tract diseases.
When digestive tract diseases occur, the characteristics of tongue, coating, tooth trace and the like often show regular changes, and the changes can be captured and analyzed by naked eyes or by an image. In modern medicine, the doctor is assisted in identifying digestive tract diseases by judging the type of digestive tract diseases corresponding to tongue images. However, in the remote consultation, because the tongue coating image shot by the mobile phone is easily affected by factors such as a mobile phone imaging algorithm, light, background interference and the like, the image cannot reflect real image information, such as white tongue coating, but the tongue coating shot by the mobile phone is reddish, and the error of tongue image recognition is caused by the deviation caused by shooting.
Disclosure of Invention
The present invention has been made in view of the above problems, and has as its object to provide a tongue image recognition method, apparatus and computer device based on image augmentation, which overcome the above problems.
In one aspect of the present invention, there is provided a tongue image recognition method based on image augmentation, the method comprising:
carrying out effective region segmentation processing on an initial tongue image to be processed to obtain a first tongue image and a first tongue outline mask image;
generating N first generated tongue images corresponding to the first tongue outline mask images based on a preset image diffusion model, wherein N is greater than or equal to 1;
respectively carrying out color correction on the first tongue images based on N first generation type tongue images, and respectively generating N second tongue images corresponding to the N first generation type tongue images one by one;
and taking N second tongue images or a set of N second tongue images and the first tongue images as target image data, and calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image.
Further, the performing color correction on the first tongue images based on the N first generated tongue images, and generating N second tongue images corresponding to the N first generated tongue images one to one respectively includes:
sequentially selecting one image from the first generated tongue image as a target generated tongue image, and performing feature point matching on the first tongue image and the target generated tongue image to obtain feature point pairs;
Generating a color correction matrix between the first tongue image and the target generated tongue image according to the color corresponding relation of the feature points;
and performing color correction on the first tongue image based on the color correction matrix to obtain a second tongue image corresponding to the target generated tongue image.
Further, the image recognition model is a cascading network model, a first network model of the cascading network model is a text description prediction model, and a second network model of the cascading network model is a tongue image classification prediction model;
the step of calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image comprises the following steps:
inputting each piece of target image data into the text description prediction model for recognition so as to obtain text descriptions corresponding to each piece of target image data, wherein the text descriptions are used for describing tongue texture features in the target image data;
and inputting text descriptions corresponding to the target image data into the tongue image classification prediction model for recognition so as to obtain a recognition result corresponding to the initial tongue image.
Further, before invoking a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image, the method further includes: pre-training the text description prediction model;
The pre-training the text description prediction model specifically comprises the following steps:
carrying out effective area segmentation processing on the first initial tongue image sample to obtain a first tongue image sample;
acquiring a text description describing tongue texture features of the first tongue image sample;
and inputting the first tongue image sample and the text description of the first tongue image sample as training data into an input layer of a preset multi-mode training model, and obtaining a text description prediction model through training.
Further, before invoking a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image, the method further includes: pre-training the tongue image classification prediction model;
the pre-training the tongue image classification prediction model specifically comprises the following steps:
carrying out effective region segmentation processing on a second initial tongue image sample to obtain a second tongue image sample and a first tongue outline mask image sample, and carrying out tongue image classification labeling on the second initial tongue image sample;
generating N first generation type tongue image samples corresponding to the second tongue outline mask image based on a preset image diffusion model;
Respectively carrying out color correction on the second tongue image samples based on N first generation type tongue image samples, and respectively generating N third tongue image samples which are in one-to-one correspondence with the N first generation type tongue image samples;
taking N third tongue image samples or a set of N third tongue image samples and second tongue image samples as target image data samples, and calling a preset text description prediction model to identify the target image data samples so as to obtain text descriptions corresponding to the target image data samples;
and inputting a tongue image classification result of the second initial tongue image sample and text descriptions corresponding to each target image data sample as training data into an input layer of a preset tongue image classification prediction training model, and obtaining tongue image classification through training.
Further, before invoking a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image, the method further includes: pre-training the tongue image classification prediction model;
the pre-training the tongue image classification prediction model specifically comprises the following steps:
acquiring text descriptions of a plurality of tongue texture features corresponding to different tongue image classification results;
And inputting any target tongue image classification result and text description of a plurality of tongue texture features corresponding to the current target tongue image classification result as training data into an input layer of a preset tongue image classification prediction training model, and obtaining the tongue image classification prediction model through training.
Further, the image recognition model is an image recognition model obtained by training in advance, and the calling the preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image comprises the following steps:
inputting each target image data into the image recognition model for recognition so as to obtain tongue image classification results corresponding to each target image data;
and carrying out statistical calculation on tongue image classification results corresponding to each target image data so as to finally obtain an initial tongue image recognition result.
Further, before the effective region segmentation processing is performed on the initial tongue image to be processed to obtain the first tongue image and the first tongue contour mask image, the method further includes: pre-training a tongue segmentation model;
the pre-trained tongue segmentation model comprises:
acquiring a third initial tongue image sample containing a tongue effective area;
Labeling the tongue effective areas in each third initial tongue image sample to obtain a semantic segmentation training sample;
inputting the semantic segmentation training sample into an input layer of a preset semantic segmentation model, and obtaining a tongue segmentation model through training.
Further, before the effective region segmentation processing is performed on the initial tongue image to be processed to obtain a first tongue image and a first tongue outline mask image, the method further comprises;
judging whether the initial tongue image contains a tongue effective area, and if not, sending out an error prompt.
In another aspect of the present invention, there is provided a tongue image recognition apparatus based on image augmentation, the apparatus comprising:
the effective area segmentation module is used for carrying out effective area segmentation processing on the initial tongue image to be processed to obtain a first tongue image and a first tongue outline mask image;
the image diffusion module is used for generating N first generated tongue images corresponding to the first tongue outline mask images based on a preset image diffusion model, wherein N is greater than or equal to 1;
the color correction module is used for respectively carrying out color correction on the first tongue images based on N first generation type tongue images and respectively generating N second tongue images which are in one-to-one correspondence with the N first generation type tongue images;
The image recognition module is used for taking N second tongue images or a set of N second tongue images and the first tongue images as target image data, and calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image.
In another aspect of the application, a computer device is provided comprising a memory, a processor, and a computer program stored on the memory and executable on the processor; the computer program, when executed by the processor, implements the steps of the tongue image recognition method based on image augmentation as described in any one of the above.
After an initial tongue image is obtained, effective area segmentation processing is carried out on the initial tongue image to obtain a first tongue image and a first tongue outline mask image; generating N first generation type tongue images corresponding to the first tongue outline mask images based on a preset image diffusion model, respectively carrying out color correction on the first tongue images based on the N first generation type tongue images, respectively generating N second tongue images corresponding to the N first generation type tongue images one by one, and improving generalization capability and robustness in image recognition by acquiring the N second tongue images after color correction; and finally, taking N second tongue images or a set of N second tongue images and the first tongue images as target image data, and calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image.
The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present invention more readily apparent.
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. In the drawings:
FIG. 1 is a flowchart of a tongue image recognition method based on image augmentation according to an embodiment of the present invention;
FIG. 2 is a schematic illustration of a tongue effective area image and a tongue contour mask of a tongue image according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a tongue image recognition device based on image augmentation 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.
It will be understood by those skilled in the art that all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs unless defined otherwise. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the prior art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
Example 1
The embodiment of the invention provides a tongue image recognition method based on image amplification, which is shown in fig. 1 and comprises the following steps:
s1, carrying out effective area segmentation processing on an initial tongue image to be processed to obtain a first tongue image and a first tongue outline mask image;
specifically, the outer contour of the tongue region can be used as an image boundary to carry out effective region segmentation, the region in the contour is used for obtaining an effective tongue image, the region outside the contour is filled with black pixels to generate a new image, the new image is cut off an ineffective region to obtain a tongue contour mask image, and part of effective characteristic information of the tongue is still reserved in the tongue contour mask image.
S2, generating N first generated tongue images corresponding to the first tongue outline mask images based on a preset image diffusion model, wherein N is greater than or equal to 1;
the image diffusion technology enables a model to learn the characteristics and the patterns of an image by training a large number of image data sets, and then generates a new generated image, wherein the generated image has the characteristics and the patterns of the learned image. The image diffusion model preset in the invention learns a large number of tongue images capable of truly reflecting tongue fur texture features, so that a first generated tongue image corresponding to a first tongue outline mask image can be generated according to the first tongue outline mask image. The first generated tongue image can truly reflect the texture characteristics of the tongue fur, and has the same structural characteristics as the first tongue image, so that the first tongue image can be used as the basis for color correction.
S3, respectively carrying out color correction on the first tongue images based on N first generation type tongue images, and respectively generating N second tongue images corresponding to the N first generation type tongue images one by one;
s4, taking N second tongue images or a set of N second tongue images and the first tongue images as target image data, and calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image.
According to the invention, the first tongue image and the first tongue outline mask image are obtained by carrying out effective area segmentation processing on the initial tongue image, so that noise interference of images outside the tongue effective area in image recognition is reduced, a plurality of first generated tongue images capable of truly reflecting the texture characteristics of the tongue are generated through the image diffusion model and the first tongue outline mask image, and as the first generated tongue image and the first tongue image are in the same position relationship, the first tongue image is subjected to color correction through the first generated tongue image, a plurality of second tongue images subjected to color correction can be obtained, and the second tongue image can truly reflect the texture characteristics of the tongue. Finally, learning and recognition are carried out based on a preset image recognition model, and finally, a recognition result of the tongue fur image picture is obtained, so that the accuracy of image recognition is improved, and misdiagnosis caused by deviation of mobile phone photography is avoided.
In an embodiment of the present invention, after obtaining the identification result, the method further includes: comparing the identification result with a pre-stored initial result, and judging whether the treatment effect is effective or not according to the comparison result; if no effect exists, pushing the initial tongue image and/or the identification result to a preset doctor system. The tongue image recognition method based on image amplification can assist doctors to finish follow-up tracking of the illness state of digestive diseases.
Specifically, the initial tongue image uploaded by the user through the mobile phone is subjected to image diffusion and color correction to obtain a plurality of tongue images after color correction, and the tongue images after color correction and the initial tongue image obtained in the original are input into the image recognition model together, so that the robustness and generalization capability of model prediction are improved, the tongue image of a patient can be accurately recognized, and a doctor can conveniently adjust a treatment scheme according to the treatment effect by checking the treatment effect of the patient at regular time.
In addition, in the embodiment of the invention, the patient can shoot and upload tongue images through the patient condition follow-up APP at the mobile phone end, and the APP can also remind the user to shoot tongue images by the mobile phone at regular time and punch cards during the medication period of the patient.
In step S1, the initial tongue image is subjected to effective region segmentation processing by using a tongue segmentation model obtained by pre-training, the tongue segmentation model uses the outer contour of the tongue region as an image boundary, the outer region of the contour is filled with black pixels to generate a new image, the new image cuts off an ineffective region, and effective characteristic information of the tongue is reserved, so that noise interference in subsequent image recognition can be reduced, and recognition accuracy is improved.
Further, before the effective region segmentation processing is performed on the initial tongue image to obtain the first tongue image and the first tongue outline mask image, the method further includes: pre-training a tongue segmentation model; the pre-trained tongue segmentation model comprises: acquiring a third initial tongue image sample containing a tongue effective area; labeling the tongue effective areas in each third initial tongue image sample to obtain a semantic segmentation training sample; inputting the semantic segmentation training sample into an input layer of a preset semantic segmentation model, and obtaining a tongue segmentation model through training.
In addition, the tongue segmentation model of the embodiment of the present invention may be further used to determine whether a tongue effective area exists in the initial tongue image, specifically, before the effective area segmentation process is performed on the initial tongue image to obtain a first tongue image and a first tongue contour mask image, the method further includes; judging whether the initial tongue image contains a tongue effective area, and if not, sending out an error prompt.
In step S2, the image diffusion model may be a Controlnet image diffusion model, which is trained based on a large number of tongue images capable of truly reflecting tongue fur texture features, and thus may be considered as a tongue image in which the first generated tongue image generated based on the image diffusion model has tongue fur texture features capable of truly reflecting tongue fur.
Further, the embodiment of the invention further includes a training method for an image diffusion model, and before generating at least one first generated tongue image corresponding to the first tongue outline mask image based on a preset image diffusion model, the method further includes: acquiring a fourth initial tongue image sample capable of truly reflecting tongue texture features; carrying out effective region segmentation processing on the fourth initial tongue image sample to obtain a fourth tongue image sample and a second tongue outline mask image sample; and inputting the fourth tongue image sample and the second tongue outline mask image sample as training data into an input layer of a preset image diffusion training model, and obtaining the image diffusion model through training.
Fig. 2 schematically illustrates a training pair of a fourth initial tongue image sample and a second tongue outline mask image according to an embodiment of the present invention, and because the tongue image in the fourth initial tongue image sample can truly reverse the texture features of the tongue, the generated tongue images generated by the trained image diffusion model in the embodiment of the present invention are tongue effective area pictures that can truly reflect the texture features of the tongue. It can be inferred that the first generated tongue images in the embodiments of the present invention are tongue effective area images that can truly reflect the tongue texture features, and the first generated tongue images can be used as references for performing color correction on the first tongue images. So as to carry out color correction on the first tongue image by referencing the color information of the first generated tongue image. The tongue texture features can be texture features of tongue quality and tongue fur.
In step S3, performing color correction on the first tongue images based on the N first generated tongue images, and generating N second tongue images corresponding to the N first generated tongue images one by one includes: sequentially selecting one image from the first generated tongue image as a target generated tongue image, and performing feature point matching on the first tongue image and the target generated tongue image to obtain feature point pairs; generating a color correction matrix between the first tongue image and the target generated tongue image according to the color corresponding relation of the feature points; and performing color correction on the first tongue image based on the color correction matrix to obtain a second tongue image corresponding to the target generated tongue image. That is, the number of second tongue images in the embodiment of the present invention is the same as the number of generated tongue images.
The color correction matrix is a technology for correcting the color deviation of the digital image, and the color correction matrix is more true and accurate by adjusting the color of the image. The basic principle is to build a matrix by measuring RGB values of a series of standard colors and RGB values of the same series of standard colors photographed by a camera, the matrix being capable of correcting color deviations of an image photographed by the camera. The matrix contains color deviation information of an image shot by the camera, and the color of each pixel point of the image can be corrected in a matrix operation mode, so that the color of the image is more real and accurate.
Further, the first generated tongue image in the embodiment of the invention has RGB values capable of truly inverting the tongue texture feature, the SIFT matching algorithm can be selected to perform feature point matching on the target generated tongue image and the first tongue image, the feature point pair is determined as an extraction key matching point, and the RGB values capable of truly reflecting the tongue texture feature are provided in the target generated tongue image, and based on certain consistency of the RGB values of the same object in different images, the color correction matrix can be calculated by matching the RGB values of the same object in different images, and the color correction can be completed by processing the first effective area image by using the color correction matrix.
According to the color correction method based on feature point matching, feature matching between two images with larger difference can be achieved, robustness and accuracy of color correction are greatly improved, and corrected images can show true colors of tongue fur.
In one embodiment of the present invention, the image recognition model in step S4 may be a cascading network model, a first network model of the cascading network model is a text description prediction model, and a second network model of the cascading network model is a tongue image classification model; the step of calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image comprises the following steps: inputting each piece of target image data into the text description prediction model for recognition so as to obtain text descriptions corresponding to each piece of target image data, wherein the text descriptions are used for describing tongue texture features in the target image data; and inputting text descriptions corresponding to the target image data into the tongue image classification model for recognition so as to obtain a recognition result corresponding to the initial tongue image.
In the embodiment of the invention, all second tongue images are used as target image data, wherein all second tongue images are tongue images after color correction. Or the second tongue image and the first tongue image are used as target image data, the original tongue image shot by the mobile phone is used as a tongue image to be learned, if the original image has no chromatic aberration, the texture characteristics of the tongue and tongue fur can be truly represented, and the generalization capability of the model can be further improved by adding the image into the model. The tongue image classification in the embodiment of the invention can classify intestinal diseases of different grades aiming at tongue image characterization, namely, based on the recognition result of the tongue image, doctors can be assisted to know the illness state of patients. It can be appreciated that in the embodiment of the present invention, tongue images may be classified according to other classification criteria, so as to accomplish different types of tongue image recognition.
Further, the embodiment of the invention also comprises a training method for the text description prediction model and the tongue image classification model. Therefore, before calling a preset image recognition model to recognize the target image data so as to obtain the recognition result of the initial tongue image, the method further comprises the following steps: pre-training the text description prediction model; the pre-training the text description prediction model specifically comprises the following steps: carrying out effective area segmentation processing on the first initial tongue image sample to obtain a first tongue image sample; acquiring a text description describing tongue texture features of the first tongue image sample; and inputting the first tongue image sample and the text description of the first tongue image sample as training data into an input layer of a preset multi-mode training model, and obtaining a text description prediction model through training. The text description of the tongue texture feature of the embodiments of the present invention can be expressed as: the tongue is pale white, the tongue shape is old and tender, the tongue coating is thick and dry, the tongue coating is white and the tongue coating is partially concentrated, and the text description of the tongue texture features is only an alternative embodiment of the invention and is not used as the basis for limiting the protection scope of the invention.
Further, before invoking a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image, the method further includes: pre-training the tongue image classification model; the pre-training the tongue image classification model specifically comprises the following steps: carrying out effective region segmentation processing on a second initial tongue image sample to obtain a second tongue image sample and a first tongue outline mask image sample, and carrying out tongue image classification labeling on the second initial tongue image sample; generating N first generation type tongue image samples corresponding to the second tongue outline mask image based on a preset image diffusion model; respectively carrying out color correction on the second tongue image samples based on N first generation type tongue image samples, and respectively generating N third tongue image samples which are in one-to-one correspondence with the N first generation type tongue image samples; taking N third tongue image samples or a set of N third tongue image samples and second tongue image samples as target image data samples, and calling a preset text description prediction model to identify the target image data samples so as to obtain text descriptions corresponding to the target image data samples; and inputting a tongue image classification result of the second initial tongue image sample and text descriptions corresponding to each target image data sample as training data into an input layer of a preset tongue image classification prediction training model, and obtaining a tongue image classification model through training. The tongue image classification result of the embodiment of the present invention may specifically be a tongue image belonging to a class C1 of chronic atrophic gastritis, a tongue image belonging to a class C2 of chronic atrophic gastritis, or the like.
Further, before invoking a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image, the method further includes: pre-training the tongue image classification model; the pre-training the tongue image classification model specifically comprises the following steps: acquiring text descriptions of a plurality of tongue texture features corresponding to different tongue image classification results; and inputting any target tongue image classification result and text description of a plurality of tongue texture features corresponding to the current target tongue image classification result as training data into an input layer of a preset tongue image classification prediction training model, and obtaining a tongue image classification model through training.
In one embodiment of the present invention, the image recognition model in step S4 may be an image recognition model, where the image recognition model is an image recognition model obtained by training in advance, and the calling a preset image recognition model to recognize the target image data, so as to obtain a recognition result of the initial tongue image, includes: inputting each target image data into the image recognition model for recognition so as to obtain tongue image classification results corresponding to each target image data; and carrying out statistical calculation on tongue image classification results corresponding to each target image data so as to finally obtain an initial tongue image recognition result. According to the embodiment, the target image data are directly subjected to learning identification based on the image identification model, after tongue image classification results corresponding to each target image data are obtained, the tongue image classification results with the largest occurrence number are counted and selected as the identification results of finally identified tongue fur image pictures.
For the purposes of simplicity of explanation, the methodologies are shown and described as a series of acts, it is to be understood and appreciated by one of ordinary skill in the art that the methodologies are not limited by the order of acts, as some acts may, in accordance with the methodologies, take place in other order or concurrently. Further, those skilled in the art will appreciate that the embodiments described in the specification are presently preferred embodiments, and that the acts are not necessarily required by the embodiments of the invention.
Example two
Fig. 3 schematically illustrates a schematic structural diagram of a tongue image recognition device based on image augmentation, and referring to fig. 3, the tongue image recognition device based on image augmentation in the embodiment of the present invention specifically includes an effective area segmentation module 301, an image diffusion module 302, a color correction module 303, and an image recognition module 304, where:
an effective area segmentation module 301, configured to perform effective area segmentation processing on an initial tongue image to be processed to obtain a first tongue image and a first tongue contour mask image;
the image diffusion module 302 is configured to generate N first generated tongue images corresponding to the first tongue contour mask images based on a preset image diffusion model, where N is greater than or equal to 1;
The color correction module 303 is configured to perform color correction on the first tongue images based on N first generated tongue images, and generate N second tongue images corresponding to the N first generated tongue images one by one;
the image recognition module 304 is configured to take N second tongue images or a set of N second tongue images and the first tongue image as target image data, and call a preset image recognition model to recognize the target image data, so as to obtain a recognition result of the initial tongue image.
Further, the color correction module 303 specifically includes:
the characteristic point matching submodule is used for sequentially selecting one image from the first generated tongue image as a target generated tongue image, and carrying out characteristic point matching on the first tongue image and the target generated tongue image to obtain a characteristic point pair;
the matrix generation sub-module is used for generating a color correction matrix between the first tongue image and the target generation type tongue image according to the color corresponding relation of the characteristic points;
and the correction calculation sub-module is used for carrying out color correction on the first tongue image based on the color correction matrix so as to obtain a second tongue image corresponding to the target generation type tongue image.
The image recognition module 304 of one embodiment of the present invention includes:
the text description prediction sub-module is used for inputting each piece of target image data into the text description prediction model for recognition so as to obtain text descriptions corresponding to each piece of target image data, wherein the text descriptions are used for describing tongue texture characteristics in the target image data;
and the tongue image classification prediction sub-module is used for inputting text descriptions corresponding to each target image data into the tongue image classification model for recognition so as to obtain a recognition result corresponding to the initial tongue image.
Further, the tongue image recognition device based on image augmentation further comprises a text description prediction model training module for training the text description prediction model in advance.
The text description prediction model training module specifically comprises:
the first acquisition submodule is used for carrying out effective area segmentation processing on the first initial tongue image sample to obtain a first tongue image sample;
a first obtaining sub-module for obtaining a text description describing tongue texture features of the first tongue image sample;
the text description training sub-module is used for inputting the first tongue image sample and the text description of the first tongue image sample as training data into an input layer of a preset multi-mode training model, and obtaining a text description prediction model through training.
Further, the tongue image recognition device based on image augmentation according to the embodiment of the invention further comprises a tongue image classification model training module for training the tongue image classification model in advance.
The tongue image classification model training module of one embodiment of the invention specifically comprises:
the second acquisition sub-module is used for carrying out effective area segmentation processing on a second initial tongue image sample to obtain a second tongue image sample and a first tongue outline mask image sample, and carrying out tongue image classification labeling on the second initial tongue image sample;
the image diffusion submodule is used for generating N first generation type tongue image samples corresponding to the second tongue outline mask image based on a preset image diffusion model;
the color correction sub-module is used for respectively carrying out color correction on the second tongue image samples based on N first generation type tongue image samples, and respectively generating N third tongue image samples which are in one-to-one correspondence with the N first generation type tongue image samples;
the text description acquisition sub-module is used for taking each third tongue image sample or a set of each third tongue image sample and the first tongue image sample as a target image data sample, and calling a preset text description prediction model to learn and identify the target image data sample so as to obtain text description corresponding to each target image data sample;
The first tongue image classification training sub-module is used for inputting a tongue image classification result of the second initial tongue image sample and text descriptions corresponding to each target image data sample as training data into an input layer of a preset tongue image classification prediction training model, and obtaining a tongue image classification model through training.
The tongue image classification model training module of another embodiment of the invention specifically comprises:
the third acquisition sub-module is used for acquiring text descriptions of a plurality of tongue texture features corresponding to different tongue image classification results;
the second tongue image classification training sub-module is used for inputting any target tongue image classification result and text description of a plurality of tongue texture features corresponding to the current target tongue image classification result as training data into an input layer of a preset tongue image classification prediction training model, and obtaining a tongue image classification model through training.
The image recognition module 304 of another embodiment of the present invention includes:
the image recognition sub-module is used for inputting each piece of target image data into the image recognition model for recognition so as to obtain tongue image classification results corresponding to each piece of target image data;
and the statistical calculation sub-module is used for carrying out statistical calculation on tongue image classification results corresponding to each target image data so as to finally obtain an initial tongue image recognition result.
Further, the tongue image recognition device based on image augmentation according to the embodiment of the invention further comprises a tongue segmentation model training module for training the tongue segmentation model in advance.
The tongue segmentation model training module specifically comprises:
a fourth acquisition sub-module for acquiring a third initial tongue image sample comprising a tongue active area;
the labeling sub-module is used for labeling the tongue effective areas in each third initial tongue image sample to obtain semantic segmentation training samples;
the semantic segmentation training sub-module is used for inputting the semantic segmentation training sample into an input layer of a preset semantic segmentation model, and obtaining a tongue segmentation model through training.
Further, the effective area segmentation module 301 of the embodiment of the present invention is further configured to determine whether the initial tongue image includes a tongue effective area, and if not, send an error prompt.
Further, the tongue image recognition device based on image augmentation according to the embodiment of the invention further comprises an image diffusion model training module for training the image diffusion model in advance.
The image diffusion model training module specifically comprises:
a fifth acquisition sub-module, configured to acquire a fourth initial tongue image sample that can truly reflect the texture characteristics of the tongue;
The effective area segmentation submodule is used for carrying out effective area segmentation processing on the fourth initial tongue image sample to obtain a fourth tongue image sample and a second tongue outline mask image sample;
and the image diffusion training sub-module is used for inputting the fourth tongue image sample and the second tongue outline mask image sample as training data into an input layer of a preset image diffusion training model, and obtaining the image diffusion model through training.
For the device embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference is made to the description of the method embodiments for relevant points.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
After an initial tongue image is obtained, a first tongue image and a first tongue outline mask image which are subjected to effective area segmentation processing on the initial tongue image are obtained; generating N first generation type tongue images corresponding to the first tongue outline mask images based on a preset image diffusion model, respectively carrying out color correction on the first tongue images based on the N first generation type tongue images, respectively generating N second tongue images corresponding to the N first generation type tongue images one by one, and improving generalization capability and robustness in image recognition by acquiring the N second tongue images after color correction; and finally, taking N second tongue images or a set of N second tongue images and the first tongue images as target image data, and calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image.
Example III
The embodiment of the invention provides a computer device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the steps in the tongue image identification method embodiments based on image augmentation, such as the steps S1-S4 shown in figure 1 when executing the computer program. Alternatively, the processor, when executing the computer program, implements the functions of the modules/units in the tongue image recognition device or gateway system embodiments based on image augmentation, such as the effective area segmentation module 301, the image diffusion module 302, the color correction module 303, and the image recognition module 304 shown in fig. 3.
Furthermore, those skilled in the art will appreciate that while some embodiments herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, any of the claimed embodiments can be used in any combination.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (11)
1. A tongue image recognition method based on image augmentation, the method comprising:
carrying out effective region segmentation processing on an initial tongue image to be processed to obtain a first tongue image and a first tongue outline mask image;
generating N first generated tongue images corresponding to the first tongue outline mask images based on a preset image diffusion model, wherein N is greater than or equal to 1;
respectively carrying out color correction on the first tongue images based on N first generation type tongue images, and respectively generating N second tongue images corresponding to the N first generation type tongue images one by one;
and taking N second tongue images or a set of N second tongue images and the first tongue images as target image data, and calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image.
2. The method according to claim 1, wherein the performing color correction on the first tongue images based on the N first generated tongue images respectively, and generating N second tongue images corresponding to the N first generated tongue images one to one respectively includes:
sequentially selecting one image from the first generated tongue image as a target generated tongue image, and performing feature point matching on the first tongue image and the target generated tongue image to obtain feature point pairs;
Generating a color correction matrix between the first tongue image and the target generated tongue image according to the color corresponding relation of the feature points;
and performing color correction on the first tongue image based on the color correction matrix to obtain a second tongue image corresponding to the target generated tongue image.
3. The method of claim 1, wherein the image recognition model is a cascading network model, a first network model of the cascading network model is a text description prediction model, and a second network model of the cascading network model is a tongue image classification prediction model;
the step of calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image comprises the following steps:
inputting each piece of target image data into the text description prediction model for recognition so as to obtain text descriptions corresponding to each piece of target image data, wherein the text descriptions are used for describing tongue texture features in the target image data;
and inputting text descriptions corresponding to the target image data into the tongue image classification prediction model for recognition so as to obtain a recognition result corresponding to the initial tongue image.
4. A method according to claim 3, wherein before invoking a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image, the method further comprises: pre-training the text description prediction model;
the pre-training the text description prediction model specifically comprises the following steps:
carrying out effective area segmentation processing on the first initial tongue image sample to obtain a first tongue image sample;
acquiring a text description describing tongue texture features of the first tongue image sample;
and inputting the first tongue image sample and the text description of the first tongue image sample as training data into an input layer of a preset multi-mode training model, and obtaining a text description prediction model through training.
5. The method of claim 4, wherein before invoking a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image, the method further comprises: pre-training the tongue image classification prediction model;
the pre-training the tongue image classification prediction model specifically comprises the following steps:
carrying out effective region segmentation processing on a second initial tongue image sample to obtain a second tongue image sample and a first tongue outline mask image sample, and carrying out tongue image classification labeling on the second initial tongue image sample;
Generating N first generation type tongue image samples corresponding to the second tongue outline mask image based on a preset image diffusion model;
respectively carrying out color correction on the second tongue image samples based on N first generation type tongue image samples, and respectively generating N third tongue image samples which are in one-to-one correspondence with the N first generation type tongue image samples;
taking N third tongue image samples or a set of N third tongue image samples and second tongue image samples as target image data samples, and calling a preset text description prediction model to identify the target image data samples so as to obtain text descriptions corresponding to the target image data samples;
and inputting a tongue image classification result of the second initial tongue image sample and text descriptions corresponding to each target image data sample as training data into an input layer of a preset tongue image classification prediction training model, and obtaining the tongue image classification prediction model through training.
6. A method according to claim 3, wherein before invoking a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image, the method further comprises: pre-training the tongue image classification prediction model;
The pre-training the tongue image classification prediction model specifically comprises the following steps:
acquiring text descriptions of a plurality of tongue texture features corresponding to different tongue image classification results;
and inputting a target tongue image classification result and text descriptions of a plurality of tongue texture features corresponding to the current target tongue image classification as training data into an input layer of a preset tongue image classification prediction training model, and obtaining the tongue image classification prediction model through training.
7. The method according to claim 1, wherein the calling a preset image recognition model to recognize the target image data to obtain a recognition result of the initial tongue image includes:
inputting each target image data into the image recognition model for recognition so as to obtain tongue image classification results corresponding to each target image data;
and carrying out statistical calculation on tongue image classification results corresponding to each target image data so as to finally obtain an initial tongue image recognition result.
8. The method according to any one of claims 1-7, wherein before performing the active area segmentation process on the initial tongue image to be processed to obtain the first tongue image and the first tongue contour mask image, the method further comprises: pre-training a tongue segmentation model;
The pre-trained tongue segmentation model comprises:
acquiring a third initial tongue image sample containing a tongue effective area;
labeling the tongue effective areas in each third initial tongue image sample to obtain a semantic segmentation training sample;
inputting the semantic segmentation training sample into an input layer of a preset semantic segmentation model, and obtaining a tongue segmentation model through training.
9. The method according to any one of claims 1-7, wherein before performing an active area segmentation process on an initial tongue image to be processed to obtain a first tongue image and a first tongue contour mask image, the method further comprises;
judging whether the initial tongue image contains a tongue effective area, and if not, sending out an error prompt.
10. A tongue image recognition device based on image augmentation, the device comprising:
the effective area segmentation module is used for carrying out effective area segmentation processing on the initial tongue image to be processed to obtain a first tongue image and a first tongue outline mask image;
the image diffusion module is used for generating N first generated tongue images corresponding to the first tongue outline mask images based on a preset image diffusion model, wherein N is greater than or equal to 1;
The color correction module is used for respectively carrying out color correction on the first tongue images based on N first generation type tongue images and respectively generating N second tongue images which are in one-to-one correspondence with the N first generation type tongue images;
the image recognition module is used for taking N second tongue images or a set of N second tongue images and the first tongue images as target image data, and calling a preset image recognition model to recognize the target image data so as to obtain a recognition result of the initial tongue image.
11. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor; the computer program, when executed by the processor, implements the tongue image recognition method based on image augmentation as claimed in any one of claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311051530.4A CN117094966B (en) | 2023-08-21 | 2023-08-21 | Tongue image identification method and device based on image amplification and computer equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311051530.4A CN117094966B (en) | 2023-08-21 | 2023-08-21 | Tongue image identification method and device based on image amplification and computer equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117094966A true CN117094966A (en) | 2023-11-21 |
| CN117094966B CN117094966B (en) | 2024-04-05 |
Family
ID=88781442
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311051530.4A Active CN117094966B (en) | 2023-08-21 | 2023-08-21 | Tongue image identification method and device based on image amplification and computer equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117094966B (en) |
Citations (41)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001186365A (en) * | 1999-12-27 | 2001-07-06 | Canon Inc | Picture processing method, picture processor and recording medium |
| KR20040059312A (en) * | 2002-12-28 | 2004-07-05 | 삼성전자주식회사 | Method of extracting tongue area from tongue image and health care service method and apparatus using tongue image |
| WO2010002070A1 (en) * | 2008-06-30 | 2010-01-07 | Korea Institute Of Oriental Medicine | Method for grouping 3d models to classify constitution |
| CN102509312A (en) * | 2011-09-20 | 2012-06-20 | 哈尔滨工业大学 | Color range space of human body digital tongue image color and extraction method thereof |
| KR20130083613A (en) * | 2012-01-13 | 2013-07-23 | 상지대학교산학협력단 | Method and apparatus for processing image of tongue using three band image |
| KR101373471B1 (en) * | 2012-10-15 | 2014-03-13 | 세종대학교산학협력단 | Apparatus and method for compensation of stereo image |
| JP2016103759A (en) * | 2014-11-28 | 2016-06-02 | 株式会社リコー | Image processing apparatus, image processing method, and program |
| US20170118407A1 (en) * | 2015-10-22 | 2017-04-27 | Samsung Electronics Co., Ltd. | Method and device for generating images |
| KR101731243B1 (en) * | 2015-12-15 | 2017-04-28 | 군산대학교 산학협력단 | A video surveillance apparatus for identification and tracking multiple moving objects with similar colors and method thereof |
| KR20170099066A (en) * | 2016-02-23 | 2017-08-31 | 정종율 | Method and system for tongue diagnosis based on image of tongue |
| US20180095533A1 (en) * | 2016-09-30 | 2018-04-05 | Samsung Electronics Co., Ltd. | Method for displaying an image and an electronic device thereof |
| CN109461128A (en) * | 2018-10-24 | 2019-03-12 | 福州大学 | A kind of consistent stereo-picture color calibration method of structure kept based on gradient |
| WO2019084919A1 (en) * | 2017-11-03 | 2019-05-09 | SZ DJI Technology Co., Ltd. | Methods and system for infrared tracking |
| CN109978873A (en) * | 2019-03-31 | 2019-07-05 | 山西慧虎健康科技有限公司 | A kind of intelligent physical examination system and method based on Chinese medicine image big data |
| WO2019137131A1 (en) * | 2018-01-10 | 2019-07-18 | Oppo广东移动通信有限公司 | Image processing method, apparatus, storage medium, and electronic device |
| WO2019218826A1 (en) * | 2018-05-17 | 2019-11-21 | 腾讯科技（深圳）有限公司 | Image processing method and device, computer apparatus, and storage medium |
| CN110598533A (en) * | 2019-07-31 | 2019-12-20 | 平安科技（深圳）有限公司 | Tongue picture matching method, electronic device, computer device, and storage medium |
| JP2020009162A (en) * | 2018-07-09 | 2020-01-16 | キヤノン株式会社 | Image processing device, image processing method and program |
| KR20200092492A (en) * | 2019-01-11 | 2020-08-04 | 연세대학교 산학협력단 | Method and Apparatus for Image Adjustment Based on Semantics-Aware |
| CN111639647A (en) * | 2020-05-22 | 2020-09-08 | 深圳市赛为智能股份有限公司 | Indicating lamp state identification method and device, computer equipment and storage medium |
| WO2020188794A1 (en) * | 2019-03-20 | 2020-09-24 | 株式会社日立国際電気 | Video system, imaging device, and video processing device |
| CN112839216A (en) * | 2021-01-13 | 2021-05-25 | 合肥埃科光电科技有限公司 | Image color correction method and device |
| CN112884682A (en) * | 2021-01-08 | 2021-06-01 | 福州大学 | Stereo image color correction method and system based on matching and fusion |
| WO2021139258A1 (en) * | 2020-06-19 | 2021-07-15 | 平安科技（深圳）有限公司 | Image recognition based cell recognition and counting method and apparatus, and computer device |
| WO2021169325A1 (en) * | 2020-02-25 | 2021-09-02 | 苏州科达科技股份有限公司 | Gaze adjustment method and apparatus, and storage medium |
| CN113724228A (en) * | 2021-08-31 | 2021-11-30 | 平安科技（深圳）有限公司 | Tongue color and coating color identification method and device, computer equipment and storage medium |
| CN113781468A (en) * | 2021-09-23 | 2021-12-10 | 河南科技大学 | Tongue image segmentation method based on lightweight convolutional neural network |
| CN113780444A (en) * | 2021-09-16 | 2021-12-10 | 平安科技（深圳）有限公司 | Tongue fur image classification model training method based on progressive learning |
| CN113781488A (en) * | 2021-08-02 | 2021-12-10 | 横琴鲸准智慧医疗科技有限公司 | Tongue picture image segmentation method, apparatus and medium |
| CN113837986A (en) * | 2020-12-15 | 2021-12-24 | 京东科技控股股份有限公司 | Method, apparatus, electronic device, and medium for recognizing tongue picture |
| WO2022033150A1 (en) * | 2020-08-11 | 2022-02-17 | Oppo广东移动通信有限公司 | Image recognition method, apparatus, electronic device, and storage medium |
| WO2022047662A1 (en) * | 2020-09-02 | 2022-03-10 | Intel Corporation | Method and system of neural network object recognition for warpable jerseys with multiple attributes |
| WO2022055037A1 (en) * | 2020-09-14 | 2022-03-17 | 안치영 | Tongue diagnostic kit for color correction and saliva examination, and non-face-to-face tongue diagnostic system using same |
| CN114372926A (en) * | 2021-12-21 | 2022-04-19 | 华东理工大学 | Traditional Chinese medicine tongue tenderness identification method based on image restoration and convolutional neural network |
| CN114708493A (en) * | 2022-02-26 | 2022-07-05 | 上海大学 | Traditional Chinese medicine crack tongue diagnosis portable device and using method |
| CN114820603A (en) * | 2022-06-27 | 2022-07-29 | 深圳中易健康科技有限公司 | Intelligent health management method based on AI tongue diagnosis image processing and related device |
| WO2022247005A1 (en) * | 2021-05-27 | 2022-12-01 | 平安科技（深圳）有限公司 | Method and apparatus for identifying target object in image, electronic device and storage medium |
| CN115965607A (en) * | 2022-12-31 | 2023-04-14 | 华东理工大学 | Intelligent traditional Chinese medicine tongue diagnosis auxiliary analysis system |
| KR20230061757A (en) * | 2021-10-29 | 2023-05-09 | 한국 한의학 연구원 | Contact type measuring apparatus for papillas of tongue and coated tongue |
| CN116187470A (en) * | 2023-01-18 | 2023-05-30 | 智慧眼科技股份有限公司 | Tongue diagnosis image color correction model training method, color correction method and equipment |
| WO2023143129A1 (en) * | 2022-01-30 | 2023-08-03 | 北京字跳网络技术有限公司 | Image processing method and apparatus, electronic device, and storage medium |
-
2023
- 2023-08-21 CN CN202311051530.4A patent/CN117094966B/en active Active
Patent Citations (42)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001186365A (en) * | 1999-12-27 | 2001-07-06 | Canon Inc | Picture processing method, picture processor and recording medium |
| KR20040059312A (en) * | 2002-12-28 | 2004-07-05 | 삼성전자주식회사 | Method of extracting tongue area from tongue image and health care service method and apparatus using tongue image |
| WO2010002070A1 (en) * | 2008-06-30 | 2010-01-07 | Korea Institute Of Oriental Medicine | Method for grouping 3d models to classify constitution |
| CN102509312A (en) * | 2011-09-20 | 2012-06-20 | 哈尔滨工业大学 | Color range space of human body digital tongue image color and extraction method thereof |
| KR20130083613A (en) * | 2012-01-13 | 2013-07-23 | 상지대학교산학협력단 | Method and apparatus for processing image of tongue using three band image |
| KR101373471B1 (en) * | 2012-10-15 | 2014-03-13 | 세종대학교산학협력단 | Apparatus and method for compensation of stereo image |
| JP2016103759A (en) * | 2014-11-28 | 2016-06-02 | 株式会社リコー | Image processing apparatus, image processing method, and program |
| US20170118407A1 (en) * | 2015-10-22 | 2017-04-27 | Samsung Electronics Co., Ltd. | Method and device for generating images |
| KR101731243B1 (en) * | 2015-12-15 | 2017-04-28 | 군산대학교 산학협력단 | A video surveillance apparatus for identification and tracking multiple moving objects with similar colors and method thereof |
| KR20170099066A (en) * | 2016-02-23 | 2017-08-31 | 정종율 | Method and system for tongue diagnosis based on image of tongue |
| US20180095533A1 (en) * | 2016-09-30 | 2018-04-05 | Samsung Electronics Co., Ltd. | Method for displaying an image and an electronic device thereof |
| WO2019084919A1 (en) * | 2017-11-03 | 2019-05-09 | SZ DJI Technology Co., Ltd. | Methods and system for infrared tracking |
| WO2019137131A1 (en) * | 2018-01-10 | 2019-07-18 | Oppo广东移动通信有限公司 | Image processing method, apparatus, storage medium, and electronic device |
| WO2019218826A1 (en) * | 2018-05-17 | 2019-11-21 | 腾讯科技（深圳）有限公司 | Image processing method and device, computer apparatus, and storage medium |
| JP2020009162A (en) * | 2018-07-09 | 2020-01-16 | キヤノン株式会社 | Image processing device, image processing method and program |
| CN109461128A (en) * | 2018-10-24 | 2019-03-12 | 福州大学 | A kind of consistent stereo-picture color calibration method of structure kept based on gradient |
| KR20200092492A (en) * | 2019-01-11 | 2020-08-04 | 연세대학교 산학협력단 | Method and Apparatus for Image Adjustment Based on Semantics-Aware |
| WO2020188794A1 (en) * | 2019-03-20 | 2020-09-24 | 株式会社日立国際電気 | Video system, imaging device, and video processing device |
| CN109978873A (en) * | 2019-03-31 | 2019-07-05 | 山西慧虎健康科技有限公司 | A kind of intelligent physical examination system and method based on Chinese medicine image big data |
| CN110598533A (en) * | 2019-07-31 | 2019-12-20 | 平安科技（深圳）有限公司 | Tongue picture matching method, electronic device, computer device, and storage medium |
| WO2021017308A1 (en) * | 2019-07-31 | 2021-02-04 | 平安科技（深圳）有限公司 | Tongue image matching method, electronic apparatus, computer device, and storage medium |
| WO2021169325A1 (en) * | 2020-02-25 | 2021-09-02 | 苏州科达科技股份有限公司 | Gaze adjustment method and apparatus, and storage medium |
| CN111639647A (en) * | 2020-05-22 | 2020-09-08 | 深圳市赛为智能股份有限公司 | Indicating lamp state identification method and device, computer equipment and storage medium |
| WO2021139258A1 (en) * | 2020-06-19 | 2021-07-15 | 平安科技（深圳）有限公司 | Image recognition based cell recognition and counting method and apparatus, and computer device |
| WO2022033150A1 (en) * | 2020-08-11 | 2022-02-17 | Oppo广东移动通信有限公司 | Image recognition method, apparatus, electronic device, and storage medium |
| WO2022047662A1 (en) * | 2020-09-02 | 2022-03-10 | Intel Corporation | Method and system of neural network object recognition for warpable jerseys with multiple attributes |
| WO2022055037A1 (en) * | 2020-09-14 | 2022-03-17 | 안치영 | Tongue diagnostic kit for color correction and saliva examination, and non-face-to-face tongue diagnostic system using same |
| CN113837986A (en) * | 2020-12-15 | 2021-12-24 | 京东科技控股股份有限公司 | Method, apparatus, electronic device, and medium for recognizing tongue picture |
| CN112884682A (en) * | 2021-01-08 | 2021-06-01 | 福州大学 | Stereo image color correction method and system based on matching and fusion |
| CN112839216A (en) * | 2021-01-13 | 2021-05-25 | 合肥埃科光电科技有限公司 | Image color correction method and device |
| WO2022247005A1 (en) * | 2021-05-27 | 2022-12-01 | 平安科技（深圳）有限公司 | Method and apparatus for identifying target object in image, electronic device and storage medium |
| CN113781488A (en) * | 2021-08-02 | 2021-12-10 | 横琴鲸准智慧医疗科技有限公司 | Tongue picture image segmentation method, apparatus and medium |
| CN113724228A (en) * | 2021-08-31 | 2021-11-30 | 平安科技（深圳）有限公司 | Tongue color and coating color identification method and device, computer equipment and storage medium |
| CN113780444A (en) * | 2021-09-16 | 2021-12-10 | 平安科技（深圳）有限公司 | Tongue fur image classification model training method based on progressive learning |
| CN113781468A (en) * | 2021-09-23 | 2021-12-10 | 河南科技大学 | Tongue image segmentation method based on lightweight convolutional neural network |
| KR20230061757A (en) * | 2021-10-29 | 2023-05-09 | 한국 한의학 연구원 | Contact type measuring apparatus for papillas of tongue and coated tongue |
| CN114372926A (en) * | 2021-12-21 | 2022-04-19 | 华东理工大学 | Traditional Chinese medicine tongue tenderness identification method based on image restoration and convolutional neural network |
| WO2023143129A1 (en) * | 2022-01-30 | 2023-08-03 | 北京字跳网络技术有限公司 | Image processing method and apparatus, electronic device, and storage medium |
| CN114708493A (en) * | 2022-02-26 | 2022-07-05 | 上海大学 | Traditional Chinese medicine crack tongue diagnosis portable device and using method |
| CN114820603A (en) * | 2022-06-27 | 2022-07-29 | 深圳中易健康科技有限公司 | Intelligent health management method based on AI tongue diagnosis image processing and related device |
| CN115965607A (en) * | 2022-12-31 | 2023-04-14 | 华东理工大学 | Intelligent traditional Chinese medicine tongue diagnosis auxiliary analysis system |
| CN116187470A (en) * | 2023-01-18 | 2023-05-30 | 智慧眼科技股份有限公司 | Tongue diagnosis image color correction model training method, color correction method and equipment |
Non-Patent Citations (2)
| Title |
|---|
| ZHUANG Q等: "Human-computer interaction based health diagnostics using ResNet34 for tongue image classification", 《COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE》, vol. 226, pages 1 - 16 * |
| 李家炜: "基于深度学习的中医舌象特征分类方法研究", 《中国优秀硕士学位论文全文数据库 医药卫生科技辑》, no. 1, pages 056 - 25 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117094966B (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110505459B (en) | Image color correction method, device and storage medium suitable for endoscope | |
| CN108960232A (en) | Model training method, device, electronic equipment and computer readable storage medium | |
| WO2021218119A1 (en) | Image toning enhancement method and method for training image toning enhancement neural network | |
| CN110502986A (en) | Identify character positions method, apparatus, computer equipment and storage medium in image | |
| CN108921161A (en) | Model training method, device, electronic equipment and computer readable storage medium | |
| CN108510560B (en) | Image processing method, image processing device, storage medium and computer equipment | |
| CN113743384B (en) | Stomach picture identification method and device | |
| CN111179252B (en) | Cloud platform-based digestive tract disease focus auxiliary identification and positive feedback system | |
| CN109360254A (en) | Image processing method and device, electronic equipment, computer readable storage medium | |
| CN111488912B (en) | Laryngeal disease diagnosis system based on deep learning neural network | |
| CN110956080A (en) | Image processing method and device, electronic equipment and storage medium | |
| CN112836653A (en) | Face privacy method, device and apparatus and computer storage medium | |
| CN110047059B (en) | Image processing method and device, electronic equipment and readable storage medium | |
| CN107920205A (en) | Image processing method, device, storage medium and electronic equipment | |
| CN112926508B (en) | Training method and device of living body detection model | |
| CN117094966B (en) | Tongue image identification method and device based on image amplification and computer equipment | |
| CN113052768B (en) | Method, terminal and computer readable storage medium for processing image | |
| CN110110750B (en) | Original picture classification method and device | |
| CN112381073A (en) | IQ (in-phase/quadrature) adjustment method and adjustment module based on AI (Artificial Intelligence) face detection | |
| CN110097080B (en) | Construction method and device of classification label | |
| CN114972065A (en) | Training method and system of color difference correction model, electronic equipment and mobile equipment | |
| CN112529002B (en) | Tongue picture classification method and device, computer equipment and storage medium | |
| CN107948619B (en) | Image processing method, device, computer readable storage medium and mobile terminal | |
| CN108881740A (en) | Image method and device, electronic equipment, computer readable storage medium | |
| CN111767829B (en) | Living body detection method, device, system and storage medium |
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 |