Traditional Chinese medicine inspection tongue diagnosis image acquisition device and method
By using standard illumination excitation components, uniform components and high-resolution cameras in traditional Chinese medicine tongue diagnosis technology, and using Snake model and clustering algorithm processing, the spectral incompleteness and spatial resolution in the existing technology are solved. The problem of low efficiency and accurate image acquisition and analysis of tongue diagnosis can be achieved.
Patent Information
- Application Number
- CN202411298317.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-18
- Publication Date
- 2024-12-27
AI Technical Summary
The existing traditional Chinese medicine tongue diagnosis technology has problems such as incomplete spectrum acquisition, low spatial resolution, long acquisition time and redundant data, which affects the accuracy and efficiency of diagnostic results.
A traditional Chinese medicine observation and tongue diagnosis image acquisition device is designed, including standard light excitation components, uniform light components and image acquisition components. Tongue diagnosis images are collected through uniform light and high-resolution cameras, and the image is performed using Snake model and clustering algorithm. Treatment, extract the characteristics of tongue coating and tongue quality.
It realizes high spatial resolution and high fidelity tongue image acquisition, improves the accuracy and efficiency of diagnosis, reduces dependence on operators, and supports quantitative analysis of traditional Chinese medicine diagnosis.
Smart Images

Figure CN119184619A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present application relate to the field of image acquisition, and more particularly to a device and method for acquiring images of digital Chinese medicine tongue diagnosis. Background Art
[0002] Tongue diagnosis in traditional Chinese medicine is a diagnostic method to understand the physiological functions and pathological changes of the human body by observing the color, shape, tongue coating and other characteristics of the patient's tongue. It is an important part of traditional Chinese medicine's visual diagnosis and is of great significance in judging the strength and weakness of the body's vital energy, the nature of the pathogenic factors, the depth of the disease, and the recovery and outcome of the disease.
[0003] With the development of technology, a variety of digital technologies have been applied to TCM tongue diagnosis to improve the objectivity, accuracy and repeatability of diagnosis. These technologies include digital imaging technology: using a high-resolution camera to capture tongue images for subsequent analysis. Spectral analysis technology: by analyzing the spectral information reflected by the tongue, a more comprehensive tongue feature can be obtained. Image processing and analysis software: algorithms are used to analyze tongue images to extract features such as color, texture, and morphology. Artificial intelligence and machine learning: AI technology is used to learn from a large amount of tongue diagnosis data to improve the accuracy and efficiency of diagnosis. Instruments are used to collect patient tongue diagnosis images.
[0004] Among the related technologies, TCM tongue diagnosis technology mainly relies on the doctor's naked eye observation, which is limited by factors such as ambient light, recognition rate, subjective feeling and clinical experience, resulting in deviations between the accurate judgment and objective evaluation of tongue images. With the development of information technology, traditional tongue diagnosis methods have been combined with modern technology to form a variety of tongue diagnosis auxiliary tools and products. Modern tongue diagnosis instruments use RGB cameras for image acquisition, but there are problems such as incomplete spectral acquisition, low spatial resolution, long acquisition time and data redundancy. Summary of the invention
[0005] The embodiments of the present application provide a device and method for acquiring images of tongue diagnosis in traditional Chinese medicine, which solves the problem that the tongue diagnosis images are not clear and thus affect the diagnosis results.
[0006] On the one hand, the present application provides a device for collecting images of traditional Chinese medicine tongue diagnosis, the device comprising:
[0007] It includes standard light excitation components, uniform light components, image acquisition components, tongue diagnosis operation table, equipment housing and cavity bracket;
[0008] A circular imaging port is provided at the collection end of the device housing, and the tongue diagnosis operation table is arranged at the circular imaging port position and is located outside the device housing; the standard light excitation component, the uniform light component and the image collection component are arranged inside the device housing;
[0009] The standard illumination excitation component is arranged at one end away from the imaging port, and excites light in the direction of the imaging port; the light homogenizing component is arranged between the quasi-illumination excitation component and the imaging port, the end of the light homogenizing component away from the imaging port surrounds the standard illumination excitation component, and the end of the light homogenizing component close to the imaging port abuts against the device housing with the circular imaging port, and evenly reflects the excitation light to the circular imaging port;
[0010] The image acquisition component is arranged on the standard light excitation component to take the tongue diagnosis image at the tongue diagnosis operation table;
[0011] The cavity support is arranged at the bottom of the device housing to support the standard light excitation component and the light homogenization component.
[0012] Specifically, the device housing includes a detachable front cover housing, a middle housing and a rear cover housing; the circular imaging port is located on the front cover housing, and the tongue diagnosis operating table is arranged on the front cover housing, including a jaw support and protective silica gel;
[0013] The protective silicone sleeve is on the circular imaging port, and the jaw support is arranged at the lower part of the circular imaging port.
[0014] Specifically, the light homogenizing assembly includes an arc-shaped trumpet-shaped light homogenizing cover and a circular light homogenizing plate, wherein the end of the arc-shaped trumpet-shaped light homogenizing cover with the largest diameter is the light emitting end, and the end with the smallest diameter is the light incident end;
[0015] The light emitting end of the light homogenizing cover is against the device housing with an imaging port, and the light incident end of the light homogenizing cover is against the standard light excitation component; the light homogenizing plate is vertically arranged inside the cover body close to the light incident end, and the diameter of the light homogenizing plate matches the diameter of the light incident end of the light homogenizing cover; the inner layers of the light homogenizing cover and the light homogenizing plate are coated with scratch-resistant diffuse reflective environmentally friendly paint for uniform light reflection; a light-transmitting hole is opened at the center of the light homogenizing plate for light transmission and camera photography.
[0016] Specifically, the cavity bracket includes a first arc-shaped clamp and a second circular clamp; the first arc-shaped clamp supports the light outlet end of the light homogenizing cover, and the standard light excitation component is installed on the second circular clamp; the light homogenizing plate is installed on the standard light excitation component.
[0017] Specifically, the diameter of the standard light excitation component matches the diameter of the second circular clamp; mounting columns are evenly penetrated in the standard light excitation component, and the light homogenizing plate is installed through the mounting columns and is arranged in parallel with the standard light excitation component.
[0018] Specifically, the standard illumination excitation assembly includes a mounting plate and light source devices evenly distributed on the mounting plate;
[0019] A mounting hole is provided at the center of the mounting plate, and the image acquisition component is fixedly installed at the mounting hole. The image acquisition component is located between the light diffuser plate and the mounting plate, facing the light transmission hole of the light diffuser plate.
[0020] Specifically, the image acquisition assembly includes a camera bracket and a camera, and the camera bracket is mounted on a mounting plate;
[0021] The camera bracket includes a base frame, at least a pair of end walls and a pair of screws. The end walls and the camera base are installed by rotating screws. The screws are movably arranged at both ends of the base frame. One of the screws is fixedly installed with the camera base, and the other screw is adjustably connected with the camera base to adjust the camera pitch angle.
[0022] Specifically, the camera resolution is not less than 12 million pixels, the optimal shooting distance is 25-30 cm, the camera lens field of view angle Vfov=27°, Hfov=35°, Dfov=43°, and the lens distortion is less than 1%.
[0023] On the other hand, the present application provides a method for standardizing the acquisition of images of traditional Chinese medicine inspection and tongue diagnosis, which is used in a device for standardizing the acquisition of images of traditional Chinese medicine inspection and tongue diagnosis, and the method comprises:
[0024] Obtain RBG tongue diagnosis images, extract tongue contours using the Snake model, and correct the extraction results;
[0025] Convert the RGB tongue diagnosis image into an image in Lab color space, and extract the tongue coating and tongue quality images through clustering algorithm;
[0026] Perform threshold segmentation on the tongue coating and tongue quality images to separate the tongue coating and tongue quality from the background;
[0027] Interpolate and smooth the extracted tongue contour to obtain a more accurate contour;
[0028] According to the contour line and threshold segmentation results, the color and contour features of the tongue coating and tongue quality are finally extracted.
[0029] Specifically, the RGB tongue diagnosis image is converted into an image in the Lab color space, and the tongue coating and the tongue coating image are extracted by a clustering algorithm, including:
[0030] Convert the RGB image to the Lab color space and extract the a and b channels;
[0031] Perform histogram equalization on channels a and b;
[0032] Merge the processed a and b channels to obtain the processed Lab image;
[0033] The Lab images were clustered using clustering algorithms, and the tongue quality and tongue coating were divided into two clusters;
[0034] According to the clustering results, the pixels belonging to the tongue quality and tongue coating in the original RGB image are marked respectively, and the tongue coating image and tongue quality image are extracted.
[0035] The beneficial effects brought by the technical solution provided in the embodiment of the present application include at least: exciting high-brightness light through a standard illumination excitation component, and then using a uniform light component connected to the standard illumination excitation component to effectively uniformly reflect the light, providing glare-free lighting to reduce shadows and reflections; and the image acquisition component is placed between the uniform light plate of the uniform light component and the mounting plate of the standard illumination excitation component, and high-definition images can be obtained in a sufficient light environment; and the image acquisition component is provided with a screw and a screw that can be adjusted to adjust the camera pitch angle to provide a full range of shooting angles. It provides a tongue image with high spatial resolution and high fidelity, which is conducive to the quantitative analysis of traditional Chinese medicine diagnosis. It has a compact structure and controllable cost, and is suitable for the field of color or characteristic wavelength image detection that requires high spatial resolution. The accuracy and efficiency of tongue diagnosis are improved through image registration and image fusion technology. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 2 is a schematic diagram of the structure of a device for collecting images of traditional Chinese medicine for visual and tongue diagnosis provided in an embodiment of the present application;
[0037] Figure 2 It is an exploded schematic diagram of a device for collecting images of traditional Chinese medicine for visual and tongue diagnosis;
[0038] Figure 3 It is a schematic diagram of the installation of the light homogenization component, cavity bracket and standard light excitation component;
[0039] Figure 4 is a schematic diagram of the structure of the cavity support;
[0040] Figure 5 is a schematic diagram of the structure of the rear cover housing 54 and the standard light excitation assembly;
[0041] Figure 6 It is a schematic diagram of the installation structure of the image acquisition component and the standard light excitation component.
[0042] Figure numerals: 10, standard light excitation component; 11, mounting column; 12, mounting plate; 13, light source device; 14, mounting hole; 20, homogenizing component; 21, homogenizing cover; 22, homogenizing plate; 23, light transmission hole; 30, image acquisition component; 31, camera bracket; 32, camera; 311, base frame; 312, end wall; 313, screw; 40, tongue diagnosis operating table; 41, jaw rest; 42, protective silicone; 50, equipment housing; 51, imaging port; 52, front cover housing; 53, middle housing; 54, rear cover housing; 60, cavity bracket; 61, first arc-shaped clamp; 62, second circular clamp; 63, power supply. DETAILED DESCRIPTION
[0043] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.
[0044] The term "multiple" as used herein refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.
[0045] Figure 1 is a schematic diagram of the structure of a device for collecting images of traditional Chinese medicine for visual and tongue diagnosis provided in an embodiment of the present application, Figure 2 It is an exploded schematic diagram of a TCM visual inspection and tongue diagnosis image acquisition device, which includes a standard light excitation component 10, a light homogenization component 20, an image acquisition component 30, a tongue diagnosis operation table 40 and a device housing 50. The standard light excitation component 10, the light homogenization component 20 and the image acquisition component 30 are arranged inside the device housing 50, and the tongue diagnosis operation table 40 is arranged outside.
[0046] The device housing 50 is a box frame as a whole, and various modules and electrical components are arranged inside, including a power supply module and a data communication module, etc. The tongue image collection end of the device housing 50 is the head end, and the end opposite to the collection end is the tail end. A circular imaging port 51 is provided at the head end of the device, and the tongue diagnosis operating table 40 is arranged at the position of the circular imaging port 51. The tongue diagnosis operating table 40 is essentially a tongue fixing device, which is convenient for obtaining clear tongue diagnosis images, such as a shooting distance of 25cm, a face and chin alignment adjustment device, an eye mask, etc. The material is selected to be easy to clean and disinfect to ensure the hygiene and safety of patients when using it.
[0047] A diffuse reflection cavity is formed inside the device housing 50, and the device housing 50 supports and protects the intermediate diffuse reflection cavity, and the standard light excitation component 10 is arranged at the rear end of the device housing 50, and excites light in the direction of the imaging port 51. In a possible embodiment, the standard light excitation component 10 can provide stable lighting that simulates natural light, such as a color temperature of 5000K or 6500K, an illuminance of 1500Lux, a color rendering index CRI greater than 95, and a spectral power distribution close to natural light with the same color temperature. Specifically, a constant current power supply is used to drive the light source to ensure the consistency of the brightness and color temperature, with a brightness error within 10Lux and a color temperature error within 50.
[0048] The light homogenizing component 20 is disposed between the standard light excitation component 10 and the imaging port 51. The end of the light homogenizing component 20 away from the imaging port 51 surrounds the standard light excitation component 10, and the end close to the imaging port 51 abuts against the device housing 50 with the circular imaging port 51, and evenly reflects the excitation light to the circular imaging port 51. The purpose of setting the light homogenizing component 20 is to provide uniform, glare-free lighting to reduce shadows and reflections.
[0049] The image acquisition component 30 is arranged at the light incident end of the light homogenizing component 20, and takes the tongue diagnosis image at the tongue diagnosis operation table 40. In some possible implementations, the image acquisition component 30 can provide a camera with high resolution, high fidelity, and high dynamic range, such as a resolution greater than 12 million pixels, an average color difference ⊿E00 of the restored standard color card less than or equal to 3, a signal-to-noise ratio SNR of a neutral gray block of a standard color card greater than 45 dB, a maximum optical distortion of the lens less than 1%, support for RAW format or high bit rate JPEG format, and ensure the richness of image data.
[0050] The device housing 50 includes a front cover housing 52, an intermediate housing 53, and a rear cover housing 54 that are interconnected. The circular imaging port 51 is located on the front cover housing 52, and the tongue diagnosis operation table 40 is arranged on the front cover housing 52, including a jaw rest 41 and protective silicone 42. The protective silicone 42 is sleeved around the circular imaging port 51, and the jaw rest 41 is arranged at the lower part of the circular imaging port 51, where the human chin is located when taking pictures. The optional jaw rest 41 can be provided with a chin alignment adjustment device for fine-tuning the position and angle.
[0051] The front cover shell 52 is the position for acquiring tongue diagnosis images, and also serves to fix the jaw rest 41. The opening is designed as an elliptical structure, so that the person being diagnosed can fully display the tongue surface position when taking tongue diagnosis photos to avoid being blocked. The excess part is blocked by the elliptical shell, and other stray light or external ambient light is prevented from entering the camera's field of view and affecting the diagnosis results.
[0052] The rear cover shell 54 fixes the light source part of the tongue diagnosis, the camera control board driver and the power supply 63 part, and has an external interface for the camera and the light source, which is convenient for connecting to a computer to control taking pictures and turning on the light source during diagnosis. In addition, a serial port of the tongue diagnosis instrument is also provided, which is a control interface for the light source and the camera, and is connected to the diagnosis computer externally to control taking pictures and turning on or off the internal light source of the tongue diagnosis instrument.
[0053] Figure 3 It is an installation diagram of a light homogenizing component, a cavity bracket and a standard light excitation component. The light homogenizing component 20 includes an arc-shaped trumpet-shaped light homogenizing cover 21 and a circular light homogenizing plate 22. In this application, the end with the largest diameter of the arc-shaped trumpet-shaped light homogenizing cover 21 is defined as the light emitting end, and the end with the smallest diameter is defined as the light incident end. The light emitting end of the light homogenizing cover 21 abuts against the device housing 50 with an imaging port 51, and the light incident end of the light homogenizing cover 21 abuts against the standard light excitation component 10. The light homogenizing plate 22 is vertically arranged inside the cover body near the light incident end, that is, near the edge of the entrance, and the diameter of the light homogenizing plate 22 matches the light entrance diameter of the light homogenizing cover 21. The inner layer of the light homogenizing cover 21 and the light homogenizing plate 22 is coated with a scratch-resistant diffuse reflective environmentally friendly paint for homogenizing and light reflection. The light homogenizing plate 22 can evenly irradiate the light on the tongue surface of the patient, and the light homogenizing cover 21 plays a secondary light homogenizing role, so that the light can be fully and evenly reflected inside. When irradiated on the tongue surface of the patient, the light uniformity of the diagnosed tongue surface is greater than 95%, and the diagnosis result is stable and reliable. In addition, a light-transmitting hole 23 is opened at the center of the light homogenizing plate 22, and the light excited by the standard light excitation component 10 enters the internal cavity through the light-transmitting hole 23. At the same time, the light-transmitting hole 23 is also used for the camera of the image acquisition component 30 to take pictures.
[0054] Figure 4 It is a structural schematic diagram of the cavity bracket, in which a cavity bracket 60 is arranged inside the intermediate shell 53, and the standard light excitation component 10 and the homogenizing component 20 are fixedly mounted on the cavity bracket 60. Specifically, the cavity bracket 60 includes a first arc-shaped clamp 61 and a second circular clamp 62, the first arc-shaped clamp 61 holds up the light outlet end of the homogenizing cover 21, the standard light excitation component 10 is mounted on the second circular clamp 62, and the homogenizing plate 22 is mounted on the standard light excitation component 10. The heights of the first arc-shaped clamp 61 and the second circular clamp 62 are different, but the axis of the fixed homogenizing cover 21 remains horizontal to the ground. Optionally, a switching power supply is also provided on the cavity bracket 60 to convert the external 220V AC power supply into a low-voltage DC power that can be used by internal cameras, light sources and other components, and to stabilize the voltage.
[0055] In some embodiments, the diameter of the standard light excitation component 10 is set to match the diameter of the second circular clamp 62, that is, the circular standard light excitation component 10 is embedded in the second circular clamp 62. In addition, the mounting posts 11 can be evenly arranged on the standard light excitation component 10, and the light homogenizing plate 22 is installed through the mounting posts 11 and is arranged parallel to the standard light excitation component 10.
[0056] In some embodiments, the standard light excitation assembly 10 includes a mounting plate 12 and a light source device 13 evenly distributed on the mounting plate 12. The light source device 13 can be designed as a ring-shaped or strip-shaped light source to reduce shadows and reflections. In addition, the light source can also be an LED light source or a xenon lamp. In this solution, the mounting plate 12 is used to fix the light source, thereby ensuring the angle consistency and uniformity of the irradiated light.
[0057] In some embodiments, the image acquisition component 30 can be mounted on the rear cover housing 54, but when the standard light excitation component 10 is far away from the rear cover housing 54, light leakage will occur, affecting the shooting quality. Therefore, in some embodiments, the image acquisition component 30 is set close to the standard light excitation component 10 or between the light homogenizing plate 22 and the mounting plate 12. The separate light source structure enables the overall structure to achieve the purpose of uniform light, and the independent unit design can ensure that the light source maintenance is more efficient and low-cost.
[0058] Figure 5 It is a schematic diagram of the structure of the rear cover shell and the standard light excitation component. Figure 61 is a schematic diagram of the installation structure of the image acquisition component and the standard light excitation component. A mounting hole 14 is provided at the center of the mounting plate 12. The image acquisition component 30 is fixedly installed at the mounting hole 14 and is located between the light diffuser 22 and the mounting plate 12, facing the light transmission hole 23 of the light diffuser 22. In particular, the image acquisition component 30 includes a camera bracket 31 and a camera 32. The camera bracket 31 is installed on the mounting plate 12. The camera bracket 31 includes a base frame 311, at least a pair of end walls 312 and a pair of screws 313. Optionally, the base frame 311 is designed as a rectangular structure, including a plurality of support legs and a fixed beam. The support legs are installed on the mounting plate 12 by bolts and other connecting parts. A protruding end wall 312 is provided on two of the oppositely arranged fixed beams, and a hole is opened on the other opposite side, and a pair of screws 313 are passed through. One of the screws 313 is fixedly installed with the base of the camera 32, and the other screw 313 is adjustable and connected with the base of the camera 32. Because the end wall 312 and the camera base are adjusted and installed by means of a screw, when the pitch angle of the camera needs to be adjusted, it is only necessary to rotate the screw and control the extension and retraction of the two screws 313 to change the pitch angle of the camera 32, so as to achieve the purpose of accurately capturing the tongue diagnosis image. In one possible implementation, the camera resolution is not less than 12 million pixels, the optimal shooting distance is 25-30 cm, the camera lens field of view angle Vfov = 27°, Hfov = 35°, Dfov = 43°, and the lens distortion is less than 1%. In some other embodiments, the camera 32 can use a 4K high-definition distortion-free camera, and in combination with a high color rendering light source, it can achieve accurate color restoration, that is, the average color difference <3, the maximum color difference <5, and high detail restoration capabilities, which fully meet the needs of tongue diagnosis.
[0059] The present application also provides an image processing and analysis module, which executes a standardized acquisition method for TCM visual inspection and tongue diagnosis images, and is used to process the photographed tongue diagnosis images, specifically including the following steps:
[0060] Obtain RBG tongue diagnosis images, extract tongue contours using the Snake model, and correct the extraction results;
[0061] Convert the RGB tongue diagnosis image into an image in Lab color space, and extract the tongue coating and tongue quality images through clustering algorithm;
[0062] Threshold segmentation is performed on the tongue coating and tongue quality images to separate them from the background;
[0063] Interpolate and smooth the extracted contour lines to obtain more accurate contour lines;
[0064] According to the contour line and threshold segmentation results, the color and contour features of the tongue coating and tongue quality are finally extracted.
[0065] The step of converting the RGB tongue diagnosis image into the Lab color space and extracting the tongue coating and tongue quality images by a clustering algorithm; this step specifically converts the RGB image into an image in the Lab color space and extracts the a and b channels;
[0066] Perform histogram equalization on channels a and b;
[0067] Merge the processed a and b channels to obtain the processed Lab image;
[0068] The Lab images were clustered using clustering algorithms, and the tongue quality and tongue coating were divided into two clusters;
[0069] According to the clustering results, the pixels belonging to the tongue quality and tongue coating in the original RGB image are marked respectively, and the adjusted tongue coating image and tongue quality image are extracted.
[0070] In the above scheme, the Snake model is an active contour model based on energy functional, which can change adaptively according to the contour shape characteristics. In the automatic extraction of tongue image features, the Snake model can be used to extract the tongue contour for subsequent feature analysis. Threshold segmentation technology is a segmentation method based on pixel grayscale values, which can divide pixels in the image with grayscale values higher or lower than a certain threshold into different regions. In the automatic extraction of tongue image features, the threshold segmentation technology can be used to binarize the tongue image for morphological processing and feature extraction.
[0071] In addition, the automatic extraction of tongue features can also be combined with other image processing and machine learning algorithms, such as edge detection, texture feature extraction, neural networks, etc., to achieve more accurate and efficient feature extraction.
[0072] The Lab color model is a color model based on human eye perception, and has a good performance on objects with complex colors such as tongue images. The colors of tongue quality and tongue coating are obviously different, so they can be separated by clustering algorithms. Commonly used clustering algorithms include K-means and DBSCAN, among which K-means is a classic clustering algorithm, and different clustering results can be obtained by adjusting the number of clusters. Histogram equalization technology can enhance the contrast of the image, making the tongue quality and tongue coating in the tongue image more obvious, which is conducive to subsequent image segmentation.
[0073] The above algorithm processing process can improve the consistency and comparability of tongue diagnosis images, automatic feature extraction and tongue texture and tongue coating separation improve the efficiency and accuracy of diagnosis, simplify the operation process, reduce the dependence on operator experience, support the electronic management of TCM tongue diagnosis data, and facilitate clinical communication and telemedicine services.
[0074] In addition to the above functions, the device system also includes image preprocessing, feature extraction, data storage and remote diagnosis functions. Image preprocessing includes color correction, denoising, enhancement, etc. to improve image quality. Feature extraction mainly involves developing algorithms to automatically identify features such as tongue color, tongue coating, and cracks. Data storage is to establish a standardized database to store tongue diagnosis images and related clinical information. Remote diagnosis technology supports network transmission of image data, which is convenient for remote expert consultation. In order to provide more accurate diagnostic results, the collected tongue images can be stored in the cloud to establish a tongue database. Through the annotation of tongue images by TCM clinical experts in TCM hospitals (including but not limited to tongue color, tongue shape, texture, etc.), a large tongue diagnosis data set is constructed, and a machine learning model is established with symptoms, pathology, treatment methods, etc., so as to cultivate digital TCM so as to realize online consultation on the patient side.
[0075] In particular, in the above data stage, the tongue pixels will be analyzed, the tongue contour will be extracted, and the tongue contour will be matched with the standard tongue contour to determine whether the tongue is distorted. This distortion is mainly caused by the sitting posture and viewing angle. When the tongue image is detected to be significantly distorted, the actual shooting distance and field of view angle range are obtained, and the distortion correction algorithm is used for correction to ensure that the acquired image is distortion-free. In addition, the system can also introduce high-fidelity image acquisition technology, which is corrected by the noise optimization algorithm within the shooting distance and field of view to ensure that the signal-to-noise ratio of the acquired image is above 45dB. Under sufficient lighting conditions, the image can also be exposed, gamma, and color optimization algorithms can be used to establish a color correction matrix to ensure the average color difference ΔE of the tongue color card. 00 Within 1, after noise reduction optimization, sharpness algorithm correction, etc., it is guaranteed that the clarity value of the ISO12233 test card is above 1000LPI.
[0076] In addition, the present application uses a uniform light component and a standard light excitation component to provide a tongue image with high spatial resolution and high fidelity, which is conducive to the quantitative analysis of TCM diagnosis. The overall structure is compact and cost-controllable, and is suitable for the field of color or characteristic wavelength image detection that requires high spatial resolution. Through image registration and image fusion technology, the accuracy and efficiency of tongue diagnosis are improved.
[0077] The preferred embodiments of the present invention are described above; it should be understood that the present invention is not limited to the above-mentioned specific embodiments, and the devices and structures not described in detail should be understood to be implemented in a common manner in the art; any technician familiar with the art can make many possible changes and modifications without departing from the technical solution of the present invention, or modify them into equivalent embodiments with equivalent changes, which does not affect the essential content of the present invention; therefore, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of protection of the technical solution of the present invention.
Claims
1. A device for collecting images of traditional Chinese medicine tongue diagnosis, characterized in that: It comprises a standard light excitation component (10), a light homogenization component (20), an image acquisition component (30), a tongue diagnosis operation table (40), a device housing (50) and a cavity support (60); A circular image-taking port (51) is provided at the collection end of the device housing (50); the tongue diagnosis operation table (40) is arranged at the position of the circular image-taking port (51) and is located outside the device housing (50); the standard light excitation component (10), the light homogenization component (20) and the image collection component (30) are arranged inside the device housing (50); The standard illumination excitation component (10) is arranged at an end away from the imaging port (51) and excites light in the direction of the imaging port (51); the light homogenizing component (20) is arranged between the quasi-illumination excitation component (10) and the imaging port (51); the end of the light homogenizing component (20) away from the imaging port (51) surrounds the standard illumination excitation component (10); the end of the light homogenizing component close to the imaging port (51) abuts against a device housing (50) having a circular imaging port (51) and evenly reflects the excitation light to the position of the circular imaging port (51); The image acquisition component (30) is arranged on the standard light excitation component (10) to capture a tongue diagnosis image at the tongue diagnosis operating table (40); The cavity support (60) is arranged at the bottom of the device housing (50) to support the standard light irradiation laser A light emitting component (10) and a light homogenizing component (20).
2. The image acquisition device for traditional Chinese medicine inspection and tongue diagnosis according to claim 1 is characterized in that: The device housing (50) comprises a detachable front housing (52), an intermediate housing (53) and a rear housing (54); the circular imaging port (51) is located on the front housing (52); the tongue diagnosis operating table (40) is arranged on the front housing (52) and comprises a jaw support (41) and protective silica gel (42); The protective silica gel (42) is sleeved on the circular imaging port (51), and the jaw support (41) is arranged at the lower part of the circular imaging port (51).
3. The image acquisition device for traditional Chinese medicine inspection and tongue diagnosis according to claim 2 is characterized in that: The light homogenizing assembly (20) comprises an arc-shaped trumpet-shaped light homogenizing cover (21) and a circular light homogenizing plate (22); the end of the arc-shaped trumpet-shaped light homogenizing cover (21) with the largest diameter is a light emitting end, and the end with the smallest diameter is a light incident end; The light emitting end of the light homogenizing cover (21) abuts against a device housing (50) provided with an imaging port (51), and the light incident end of the light homogenizing cover (21) abuts against the standard light excitation component (10); the light homogenizing plate (22) is vertically arranged inside the cover body close to the light incident end, and the diameter of the light homogenizing plate (22) matches the diameter of the light incident end of the light homogenizing cover (21); the inner layers of the light homogenizing cover (21) and the light homogenizing plate (22) are coated with a scratch-resistant diffuse reflective environmentally friendly coating for uniform light reflection; a light-transmitting hole (23) is provided at the center of the light homogenizing plate (22) for light transmission and camera photography.
4. The image acquisition device for traditional Chinese medicine inspection and tongue diagnosis according to claim 1, characterized in that: The cavity support (60) comprises a first arc-shaped clamp (61) and a second circular clamp (62); the first arc-shaped clamp (61) holds up the light outlet end of the light homogenizing cover (21), and the standard light excitation component (10) is mounted on the second circular clamp (62); and the light homogenizing plate (22) is mounted on the standard light excitation component (10).
5. The image acquisition device for traditional Chinese medicine inspection and tongue diagnosis according to claim 4 is characterized in that: The diameter of the standard light excitation component (10) matches the diameter of the second circular clamp (62); the standard light excitation component (10) is evenly penetrated with mounting columns (11), and the light homogenizing plate (22) is installed through the mounting columns (11) and is arranged in parallel with the standard light excitation component (10).
6. The image acquisition device for traditional Chinese medicine inspection and tongue diagnosis according to claim 3, characterized in that: The standard light excitation assembly (10) comprises a mounting plate (12) and light source devices (13) evenly distributed on the mounting plate (12); A mounting hole (14) is provided at the center of the mounting plate (12); the image acquisition component (30) is fixedly mounted at the mounting hole (14); and the image acquisition component (30) is located between the light diffuser (22) and the mounting plate (12), facing the light transmission hole (23) of the light diffuser (22).
7. The image standardization acquisition device for TCM visual inspection and tongue diagnosis according to claim 6 is characterized in that: The image acquisition component (30) comprises a camera bracket (31) and a camera (32), wherein the camera bracket (31) is mounted on a mounting plate (12); The camera support (31) comprises a base frame (311), at least a pair of end walls (312) and a pair of screw rods (313); the end walls (312) and the base of the camera (32) are mounted via screws; the screw rods (313) are movably arranged at both ends of the base frame (311); one of the screw rods (313) is fixedly mounted on the base of the camera (32); and the other screw rod (313) is adjustably connected to the base of the camera (32) for adjusting the camera pitch angle.
8. The image standardization acquisition device for TCM visual inspection and tongue diagnosis according to claim 7 is characterized in that: The camera resolution is no less than 12 million pixels, the optimal shooting distance is 25-30 cm, the camera lens field of view angle Vfov = 27°, Hfov = 35°, Dfov = 43°, and the lens distortion is less than 1%.
9. A method for standardizing the acquisition of images of traditional Chinese medicine inspection and tongue diagnosis, used in the device for standardizing the acquisition of images of traditional Chinese medicine inspection and tongue diagnosis as claimed in any one of claims 1 to 8, characterized in that: The method comprises: Obtain RBG tongue diagnosis images, extract tongue contours using the Snake model, and correct the extraction results; Convert the RGB tongue diagnosis image into an image in Lab color space, and extract the tongue coating and tongue quality images through clustering algorithm; Perform threshold segmentation on the tongue coating and tongue quality images to separate the tongue coating and tongue quality from the background; Interpolate and smooth the extracted tongue contour to obtain a more accurate contour; According to the contour line and threshold segmentation results, the color and contour features of the tongue coating and tongue quality are finally extracted.
10. The method according to claim 9, characterized in that The step of converting the RGB tongue diagnosis image into an image in the Lab color space and extracting the tongue coating and the tongue coating image by a clustering algorithm includes: Convert the RGB image to the Lab color space and extract the a and b channels; Perform histogram equalization on channels a and b; Merge the processed a and b channels to obtain the processed Lab image; The Lab images were clustered using clustering algorithms, and the tongue quality and tongue coating were divided into two clusters; According to the clustering results, the pixels belonging to the tongue quality and tongue coating in the original RGB image are marked respectively, and the tongue coating image and tongue quality image are extracted.
