Disease diagnosis method and related device

Abstract

The invention discloses a disease diagnosis method and a related device. The disease diagnosis method comprises the following steps: acquiring a plurality of confirmed first tongue imaging images of afirst patient and first semantic segmentation images corresponding to the first tongue imaging images; determining a category label of each of a plurality of regions included in the first semantic segmentation image; Inputting the first tongue imaging image, the first semantic segmentation image and the category label into a to-be-trained model for training to obtain an image semantic segmentation model; And inputting the second tongue imaging image of the second patient to be inquired into the image semantic segmentation model to obtain a second semantic segmentation image corresponding tothe second tongue imaging image, and determining the disease type of the second patient according to the second semantic segmentation image. By means of the tongue diagnosis method and device, the area where the tongue body is located can be accurately separated from the tongue imaging image, interference of other image contents such as the face and the lip on the tongue diagnosis result is reduced, and the tongue diagnosis accuracy is improved.

Description

technical field [0001] This application relates to the field of machine learning, in particular to a disease diagnosis method and related equipment. Background technique [0002] In order to allow ordinary people to use traditional Chinese medical techniques to diagnose diseases at home and reduce the workload of doctors, the artificial intelligence tongue diagnosis application (Application, App) based on mobile clients has been extensively studied. Only need to input the patient's tongue imaging image into Tongue Diagnosis App to diagnose the patient's disease type, such as digestive system disease, respiratory system disease, etc. Artificial intelligence tongue diagnosis technology is especially suitable for detecting diseases for children aged 0-16, because the tongue image of children in this age group has fewer interference factors than adults, and it is easier to diagnose the real disease condition. [0003] Among the existing artificial intelligence tongue diagnosis ...

AI Technical Summary

Problems solved by technology

[0003] Among the existing artificial intelligence tongue diagnosis technologies, some use traditional image positioning technology to locate the tongue in images uploaded by users, such as harr, snake, etc. However, the robustness of traditional image positioning technology is poor, if the tongue in the image is If the shape or color of the tongue changes a lot, it may not be able to locate it well

The other part is to use fixed instruments to take imaging images of the patient's tongue in a fixed environment. Although the effect is good, the cost is too high, which is not conducive to popularization

There is also a part that uses the method of target detection to determine the area where the tongue is located and the determined area is usually rectangular. However, the tongue is not rectangular. Therefore, if only a rectangular frame is used to locate the tongue, and the tongue is based on the image in the rectangular frame diagnosis, the noise in this box will seriously affect the accuracy of the diagnosis result

Images