Tongue picture abnormal feature detection method and device, electronic equipment and product
By using a two-stage detection model to crop and magnify the tongue region, the problem of low accuracy in detecting abnormal tongue features was solved, and accurate identification of microscopic features such as cracks and tooth marks was achieved, thus improving the reliability of the detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHENGDU ZIJIELIU TECH CO LTD
- Filing Date
- 2025-12-24
- Publication Date
- 2026-04-17
AI Technical Summary
The accuracy of tongue abnormality detection in existing technologies is low, mainly because single-stage target detection models have a high false negative rate and are significantly affected by background noise when processing tongue localization and microscopic feature detection.
A two-stage detection model is adopted. First, the tongue region is detected by the first target detection model. Then, the tongue region is cropped and linearly interpolated to enlarge it. Finally, the abnormal feature is detected by the second target detection model to avoid background noise interference and improve resolution.
It improves the accuracy of detecting abnormal tongue features, accurately identifies microscopic features such as cracks and tooth marks, and provides reliable detection basis.
Smart Images
Figure CN121883385A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of target detection technology, specifically relating to a method, device, electronic equipment, and product for detecting abnormal tongue features. Background Technology
[0002] Tongue examination is one of the core methods of TCM diagnostics. By observing the shape, color, moisture, and changes in the tongue coating, one can help determine the health status of the human body. Among them, fissured tongue and scalloped tongue are two types of abnormal tongue appearances with significant clinical significance.
[0003] With the development of artificial intelligence technology, machine learning techniques for detecting cracks and teeth marks on the tongue are widely used to assist in traditional Chinese medicine diagnosis. Currently, the most common method for detecting cracks and teeth marks on the tongue is to directly identify abnormal features such as cracks and teeth marks using a single-stage target detection model. However, when identifying abnormal features such as cracks and teeth marks using a single-stage target detection model, the simultaneous processing of tongue localization and microscopic feature detection leads to a high rate of missed detection for small target features such as cracks and teeth marks. Furthermore, the background noise in the original image can affect the model's judgment of subtle features, resulting in a low accuracy rate for detecting abnormal features on the tongue.
[0004] Therefore, how to provide an effective solution to improve the accuracy of tongue abnormality detection has become a pressing problem in existing technologies. Summary of the Invention
[0005] The purpose of this invention is to provide a method, device, electronic device, and product for detecting abnormal tongue features, in order to solve the above-mentioned problems existing in the prior art.
[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for detecting abnormal tongue features, comprising: The panoramic image of the tongue surface is input into a pre-trained first target detection model to detect the tongue region, thereby obtaining the tongue region in the panoramic image of the tongue surface. The tongue region in the panoramic image of the tongue surface is cropped to obtain a cropped image; After the cropped image is linearly interpolated and enlarged, it is input into a pre-trained second target detection model for anomaly feature detection to obtain anomaly feature detection results. The abnormal feature detection results include the coordinates of the abnormal features and their confidence levels, and the abnormal features include cracks and / or tooth marks.
[0007] In one possible design, the method further includes: Scale the panoramic image of the tongue surface to a first specified size; The step of inputting a panoramic image of the tongue surface into a pre-trained first target detection model for tongue region detection includes: The panoramic image of the tongue surface, scaled to a first specified size, is input into the first target detection model for tongue region detection.
[0008] In one possible design, the resolution of the panoramic image of the tongue surface scaled to a first specified size is 640×640.
[0009] In one possible design, the cropped image is linearly interpolated and enlarged, including: The cropped image is enlarged to a second specified size using bilinear interpolation.
[0010] In one possible design, the resolution of the cropped image after linear interpolation is 640×640.
[0011] In one possible design, both the first object detection model and the second object detection model are YOLOv11 models.
[0012] In one possible design, before inputting the panoramic image of the tongue surface into a pre-trained first target detection model for tongue region detection, the method further includes: Obtain a first training sample set and a second training sample set. The first training sample set includes multiple panoramic images of the tongue surface of the samples, and the second training sample set includes multiple images of the tongue region of the samples after linear interpolation and magnification. The sample tongue panoramic image in the first training sample set is used as the sample input of the first target detection model, and the tongue region in the sample tongue panoramic image is used as the sample output for training, so as to obtain the trained first target detection model. The tongue region images from the second training sample set are used as the input samples for the second target detection model, and the coordinates and confidence scores of abnormal features in the tongue region images are used as the output samples for training, thus obtaining the trained second target detection model.
[0013] Secondly, the present invention provides a device for detecting abnormal tongue features, comprising: The first detection unit is used to input the panoramic image of the tongue surface into a pre-trained first target detection model to detect the tongue region and obtain the tongue region in the panoramic image of the tongue surface. The cropping unit is used to crop the tongue region in the panoramic image of the tongue surface to obtain a cropped image; The second detection unit is used to enlarge the cropped image by linear interpolation and then input it into a pre-trained second target detection model to detect abnormal features and obtain abnormal feature detection results. The abnormal feature detection results include the coordinates of the abnormal features and their confidence levels, and the abnormal features include cracks and / or tooth marks.
[0014] Thirdly, the present invention provides an electronic device comprising a memory, a processor, and a transceiver connected in sequence and communication, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the tongue abnormality feature detection method as described in the first aspect or any possible design of the first aspect.
[0015] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, perform the tongue abnormality feature detection method described in the first aspect or any possible design of the first aspect.
[0016] Fifthly, the present invention provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the tongue abnormality feature detection method as described in the first aspect or any possible design of the first aspect.
[0017] Beneficial effects: This invention establishes a first target detection model and a second target detection model. The first target detection model is used to detect the tongue region, enabling macroscopic feature localization of the tongue region. The located tongue region is then cropped and amplified using linear interpolation before being used for anomaly feature detection via the second target detection model. This avoids interference from background noise in anomaly feature detection. Simultaneously, linear interpolation amplification preserves original pixel information while achieving high-resolution input, preventing distortion of microscopic features. This allows for accurate identification of microscopic features such as cracks and tooth marks, improving the accuracy of tongue anomaly feature detection and providing a reliable basis for tongue image detection, facilitating practical application and promotion. Attached Figure Description
[0018] Figure 1 A flowchart of the tongue abnormality feature detection method provided in the embodiments of this application; Figure 2 A block diagram of the tongue abnormality feature detection device provided in the embodiments of this application; Figure 3 This is a block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation
[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.
[0020] It should be understood that although the terms "first" and "second", etc., may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another. For example, the first object may be referred to as the second object, and similarly, the second object may be referred to as the first object, without departing from the scope of the exemplary embodiments of the invention.
[0021] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, or A and B exist simultaneously. Another example is A, B and / or C, which can mean that any one of A, B, and C or any combination thereof exists. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone or A and B exist simultaneously. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.
[0022] To improve the accuracy of detecting abnormal tongue features, this application provides a method, device, electronic device, and product for detecting abnormal tongue features. This method, device, electronic device, and product can avoid the interference of background noise on the detection of abnormal features and avoid microscopic feature distortion, thereby improving the accuracy of detecting abnormal tongue features.
[0023] like Figure 1 As shown, the first aspect of the present application provides a method for detecting abnormal tongue features, which may include, but is not limited to, the following steps S101-S103.
[0024] Step S101. Input the panoramic image of the tongue surface into the pre-trained first target detection model to detect the tongue region and obtain the tongue region in the panoramic image of the tongue surface.
[0025] Among them, the panoramic image of the tongue can include the captured image of the tongue surface and its surrounding area.
[0026] When detecting abnormal tongue features such as cracks and teeth marks, a panoramic image of the tongue surface can be input into a pre-trained first target detection model to detect the tongue region and obtain the tongue region in the panoramic image of the tongue surface.
[0027] In one or more embodiments, the first target detection model can be trained using a sample panoramic image of the tongue surface as input and the tongue region in the sample panoramic image as output. Specifically, when training the first target detection model, a first training sample set can be obtained, which includes multiple sample panoramic images of the tongue surface. Then, each sample panoramic image of the tongue surface in the first training sample set is used as the input of the first target detection model, and the tongue region in each sample panoramic image is used as the output to train the model, thus obtaining the trained first target detection model.
[0028] The first object detection model may, but is not limited to, the YOLO (You Only Look Once) series of models or the SSD (Single Shot MultiBox Detector) model. In one or more embodiments, the first object detection model adopts the YOLOv11 model.
[0029] In one or more embodiments, before performing tongue region detection, the panoramic image of the tongue surface can be scaled to a first specified size, and then the scaled panoramic image of the tongue surface is input into the first target detection model for tongue region detection. The resolution of the scaled panoramic image of the tongue surface can be determined according to the selected model. For example, when the first target detection model is the YOLOv11 model, the resolution of the panoramic image of the tongue surface scaled to the first specified size can be 640×640.
[0030] Step S102. Cropping the tongue region in the panoramic image of the tongue surface to obtain a cropped image.
[0031] By cropping the tongue region in the panoramic image of the tongue, the background image outside the tongue region can be removed, thus avoiding interference from background noise on subsequent abnormal feature detection.
[0032] In one or more embodiments, when cropping the tongue region in a panoramic image of the tongue surface, the cropping size can be determined based on the size of the image input to the second object detection model, ensuring as much as possible that the length and width of the image input to the second object detection model are exactly positive integer multiples of the length and width of the cropped image.
[0033] Step S103. After the cropped image is enlarged by linear interpolation, it is input into the pre-trained second target detection model for anomaly feature detection to obtain the anomaly feature detection result.
[0034] The abnormal feature detection results include the coordinates of the abnormal features and their confidence levels, and the abnormal features include cracks and / or tooth marks.
[0035] In one or more embodiments, the second target detection model can be trained using linearly interpolated enlarged sample tongue region images as input and the coordinates and confidence scores of abnormal features in the sample tongue region images as output. Specifically, when training the second target detection model, a second training sample set can be obtained first. The second training sample set includes multiple linearly interpolated enlarged sample tongue region images. Then, the sample tongue region images in the second training sample set are used as input to the second target detection model, and the coordinates and confidence scores of abnormal features in the sample tongue region images are used as output to train the model, thus obtaining the trained second target detection model.
[0036] The second object detection model may, but is not limited to, the YOLO (You Only Look Once) series of models or the SSD (Single Shot MultiBox Detector) model. In one or more embodiments, the second object detection model adopts the YOLOv11 model.
[0037] When performing linear interpolation to enlarge the cropped image, the cropped image can be enlarged to a second specified size through, but is not limited to, bilinear interpolation, trilinear interpolation, or multilinear interpolation. The resolution of the cropped image after linear interpolation enlargement can be determined according to the selected model. For example, when the second object detection model is the YOLOv11 model, the resolution of the cropped image after linear interpolation enlargement can be 640×640.
[0038] This invention establishes a first target detection model and a second target detection model. The first target detection model is used to detect the tongue region, enabling macroscopic feature localization of the tongue region. The located tongue region is then cropped and amplified using linear interpolation before being used for anomaly feature detection via the second target detection model. This avoids interference from background noise in anomaly feature detection. Simultaneously, linear interpolation amplification preserves original pixel information while achieving high-resolution input, preventing distortion of microscopic features. This allows for accurate identification of microscopic features such as cracks and tooth marks, improving the accuracy of tongue anomaly feature detection and providing a reliable basis for tongue image detection, facilitating practical application and promotion.
[0039] Please see Figure 2 The second aspect of this application provides a tongue abnormality feature detection device, which includes: The first detection unit is used to input the panoramic image of the tongue surface into a pre-trained first target detection model to detect the tongue region and obtain the tongue region in the panoramic image of the tongue surface. The cropping unit is used to crop the tongue region in the panoramic image of the tongue surface to obtain a cropped image; The second detection unit is used to enlarge the cropped image by linear interpolation and then input it into a pre-trained second target detection model to detect abnormal features and obtain abnormal feature detection results. The abnormal feature detection results include the coordinates of the abnormal features and their confidence levels, and the abnormal features include cracks and / or tooth marks.
[0040] The working process, working details and technical effects of the tongue abnormality feature detection device provided in the second aspect of this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.
[0041] like Figure 3 As shown, a third aspect of this application provides an electronic device, including a memory, a processor, and a transceiver that are sequentially and communicatively connected, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the tongue abnormality feature detection method as described in the first aspect of the embodiment.
[0042] Specifically, the memory may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or last-in-first-out (FILO) memory, etc.; the processor may not be limited to microprocessors of the STM32F105 series, ARM (Advanced RISC Machines), x86 architecture processors, or processors with integrated NPU (neural-network processing units); the transceiver may be, but is not limited to, WiFi (Wireless Fidelity) wireless transceivers, Bluetooth wireless transceivers, General Packet Radio Service (GPRS) wireless transceivers, ZigBee (a low-power LAN protocol based on the IEEE 802.15.4 standard), 3G transceivers, 4G transceivers, and / or 5G transceivers, etc.
[0043] This fourth aspect of the embodiment provides a computer-readable storage medium storing instructions comprising the tongue abnormality feature detection method described in the first aspect of the embodiment. Specifically, the computer-readable storage medium stores instructions that, when executed on a computer, perform the tongue abnormality feature detection method as described in the first aspect. The computer-readable storage medium refers to a data storage medium, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
[0044] The fifth aspect of this embodiment provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the tongue abnormality feature detection method as described in the first aspect of the embodiment, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.
[0045] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for detecting abnormal tongue features, characterized in that, include: The panoramic image of the tongue surface is input into a pre-trained first target detection model to detect the tongue region, thereby obtaining the tongue region in the panoramic image of the tongue surface. The tongue region in the panoramic image of the tongue surface is cropped to obtain a cropped image; After the cropped image is linearly interpolated and enlarged, it is input into a pre-trained second target detection model for anomaly feature detection to obtain anomaly feature detection results. The abnormal feature detection results include the coordinates of the abnormal features and their confidence levels, and the abnormal features include cracks and / or tooth marks.
2. The method for detecting abnormal tongue features according to claim 1, characterized in that, The method further includes: Scale the panoramic image of the tongue surface to a first specified size; The step of inputting a panoramic image of the tongue surface into a pre-trained first target detection model for tongue region detection includes: The panoramic image of the tongue surface, scaled to a first specified size, is input into the first target detection model for tongue region detection.
3. The method for detecting abnormal tongue features according to claim 2, characterized in that, The resolution of the panoramic image of the tongue surface, scaled to the first specified size, is 640×640.
4. The method for detecting abnormal tongue features according to claim 1, characterized in that, The cropped image is then enlarged using linear interpolation, including: The cropped image is enlarged to a second specified size using bilinear interpolation.
5. The method for detecting abnormal tongue features according to claim 1, characterized in that, The resolution of the cropped image after linear interpolation is 640×640.
6. The method for detecting abnormal tongue features according to claim 1, characterized in that, Both the first target detection model and the second target detection model are YOLOv11 models.
7. The method for detecting abnormal tongue features according to claim 1, characterized in that, Before inputting the panoramic image of the tongue surface into a pre-trained first target detection model for tongue region detection, the method further includes: Obtain a first training sample set and a second training sample set. The first training sample set includes multiple panoramic images of the tongue surface of the samples, and the second training sample set includes multiple images of the tongue region of the samples after linear interpolation and magnification. The sample tongue panoramic image in the first training sample set is used as the sample input of the first target detection model, and the tongue region in the sample tongue panoramic image is used as the sample output for training, so as to obtain the trained first target detection model. The tongue region images from the second training sample set are used as the input samples for the second target detection model, and the coordinates and confidence scores of abnormal features in the tongue region images are used as the output samples for training, thus obtaining the trained second target detection model.
8. A device for detecting abnormal tongue features, characterized in that, include: The first detection unit is used to input the panoramic image of the tongue surface into a pre-trained first target detection model to detect the tongue region and obtain the tongue region in the panoramic image of the tongue surface. The cropping unit is used to crop the tongue region in the panoramic image of the tongue surface to obtain a cropped image; The second detection unit is used to enlarge the cropped image by linear interpolation and then input it into a pre-trained second target detection model to detect abnormal features and obtain abnormal feature detection results. The abnormal feature detection results include the coordinates of the abnormal features and their confidence levels, and the abnormal features include cracks and / or tooth marks.
9. An electronic device, characterized in that, The device includes a memory, a processor, and a transceiver that are sequentially and communicatively connected. The memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the tongue abnormality feature detection method as described in any one of claims 1 to 7.
10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement the tongue abnormality feature detection method as described in any one of claims 1 to 7.