CN-122025070-A - Intelligent skin disease inquiry method, intelligent skin disease inquiry system, electronic equipment and storage medium

CN122025070ACN 122025070 ACN122025070 ACN 122025070ACN-122025070-A

Abstract

The invention discloses an intelligent skin disease inquiry method, an intelligent skin disease inquiry system, electronic equipment and a storage medium, which relate to the field of intelligent skin disease inquiry, wherein the method comprises the steps of obtaining skin disease pictures and disease texts of patients; the method comprises the steps of carrying out feature coding according to a multi-mode model according to a skin disease picture and a disease text to obtain an image feature vector and a text feature vector, carrying out similar image matching according to the image feature vector and a skin disease image database to obtain a similar image and corresponding text information, determining a problem library text feature vector according to the text information corresponding to the similar image by using a problem library, and determining skin disease type and probability according to the problem library text feature vector and the text feature vector by using a skin disease classification model. The invention can improve the efficiency of on-line inquiry.

Inventors

LI HUANYU
WEI ZIKUN
ZHU LING
LV JUNWEI

Assignees

云南云科特色植物提取实验室有限公司
上海贝芙汀科技有限公司
李寰宇

Dates

Publication Date: 20260512
Application Date: 20231024

Claims (10)

1. An intelligent skin disease inquiry method is characterized by comprising the following steps: obtaining a dermatological picture and a symptom text of a patient; Performing feature coding by utilizing a multi-mode model according to the dermatological picture and the disorder text to obtain an image feature vector and a text feature vector; performing similar image matching according to the image feature vector and a skin disease image database to obtain a similar image and corresponding text information; determining a text feature vector of a question bank by using a question bank according to the text information corresponding to the similar images; And determining the skin disease type and probability by using a skin disease classification model according to the text feature vector of the question bank and the text feature vector.
2. The intelligent skin disease inquiry method according to claim 1, wherein the matching of similar images is performed according to the image feature vector and a skin disease image database to obtain similar images and corresponding text information, and the method specifically comprises the following steps: Calculating cosine similarity distance between the image feature vector and the image feature vector in the skin disease image database; and determining similar images and corresponding text information according to the cosine similar distance.
3. The intelligent skin disease inquiry method according to claim 1, wherein the determining the text feature vector of the question bank by using the question bank according to the text information corresponding to the similar image specifically comprises: question asking is carried out by utilizing questions in a question bank according to text information corresponding to the similar images, so that combined rule phrases are obtained; and determining the text feature vector of the question bank by utilizing the multi-modal model according to the combined rule phrase.
4. The intelligent skin condition inquiry method according to claim 1, wherein the skin condition type and probability are determined by using a skin condition classification model according to the question bank text feature vector and the text feature vector, and specifically comprising: Combining the text feature vector of the question bank with the text feature vector to obtain a combined text vector; And determining the type and probability of the skin disease by using a skin disease classification model according to the merged text vector and the image feature vector, wherein the skin disease classification model is a support vector machine.
5. An intelligent skin condition inquiry system, comprising: the acquisition module is used for acquiring skin disease pictures and disorder texts of the patient; The feature coding module is used for carrying out feature coding by utilizing a multi-mode model according to the skin disease picture and the disease text to obtain an image feature vector and a text feature vector; the similar image matching module is used for carrying out similar image matching according to the image feature vector and the skin disease image database to obtain a similar image and corresponding text information; The question library question module is used for determining text feature vectors of the question library by using the question library according to the text information corresponding to the similar images; And the skin disease determining module is used for determining skin disease types and probabilities by using a skin disease classification model according to the text feature vector of the question bank and the text feature vector.
6. The intelligent skin condition inquiry system according to claim 5, wherein the similar image matching module specifically comprises: A calculating unit for calculating cosine similarity distance between the image feature vector and the image feature vector in the skin disease image database; and the matching unit is used for determining similar images and corresponding text information according to the cosine similar distance.
7. The intelligent skin disease inquiry system of claim 5, wherein the question bank inquiry module specifically comprises: the questioning unit is used for questioning by utilizing questions in a question bank according to the text information corresponding to the similar images to obtain combined rule phrases; And the question bank text feature determining unit is used for determining a question bank text feature vector by utilizing a multi-mode model according to the combined rule phrases.
8. The intelligent skin condition interrogation system of claim 5, wherein the skin condition determination module comprises: The merging unit is used for merging the text features of the question library and the text feature vectors to obtain merged text vectors; And the skin disease determining unit is used for determining skin disease types and probabilities by utilizing a skin disease classification model according to the merged text vector and the image feature vector, wherein the skin disease classification model is a support vector machine.
9. An electronic device, comprising: one or more processors; a storage device having one or more programs stored thereon; the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the method of any of claims 1-4.
10. A computer storage medium, having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the method of any of claims 1 to 4.

Description

Intelligent skin disease inquiry method, intelligent skin disease inquiry system, electronic equipment and storage medium Technical Field The invention relates to the field of intelligent inquiry, in particular to an intelligent inquiry method, an intelligent inquiry system, electronic equipment and a storage medium for skin diseases. Background Most skin diseases have significant features on the image, and doctors can diagnose patients directly through an on-line inquiry mode. The on-line inquiry greatly simplifies the medical treatment process and saves the time of doctors and patients. But the on-line inquiry efficiency is low at present. Disclosure of Invention The invention aims to provide an intelligent skin disease inquiry method, an intelligent skin disease inquiry system, electronic equipment and a storage medium, which can improve the efficiency of online inquiry. In order to achieve the above object, the present invention provides the following solutions: An intelligent skin disease inquiry method comprising: obtaining a dermatological picture and a symptom text of a patient; Performing feature coding by utilizing a multi-mode model according to the dermatological picture and the disorder text to obtain an image feature vector and a text feature vector; performing similar image matching according to the image feature vector and a skin disease image database to obtain a similar image and corresponding text information; determining a text feature vector of a question bank by using a question bank according to the text information corresponding to the similar images; And determining the skin disease type and probability by using a skin disease classification model according to the text feature vector of the question bank and the text feature vector. Optionally, performing similar image matching according to the image feature vector and the dermatological image database to obtain a similar image and corresponding text information, which specifically includes: Calculating cosine similarity distance between the image feature vector and the image feature vector in the skin disease image database; and determining similar images and corresponding text information according to the cosine similar distance. Optionally, determining the text feature vector of the question bank by using the question bank according to the text information corresponding to the similar image specifically includes: question asking is carried out by utilizing questions in a question bank according to text information corresponding to the similar images, so that combined rule phrases are obtained; and determining the text feature vector of the question bank by utilizing the multi-modal model according to the combined rule phrase. Optionally, determining the skin disease type and probability by using a skin disease classification model according to the text feature vector of the question bank and the text feature vector specifically includes: Combining the text feature vector of the question bank with the text feature vector to obtain a combined text vector; And determining the type and probability of the skin disease by using a skin disease classification model according to the merged text vector and the image feature vector, wherein the skin disease classification model is a support vector machine. The invention also provides an intelligent skin disease inquiry system, which comprises: the acquisition module is used for acquiring skin disease pictures and disorder texts of the patient; The feature coding module is used for carrying out feature coding by utilizing a multi-mode model according to the skin disease picture and the disease text to obtain an image feature vector and a text feature vector; the similar image matching module is used for carrying out similar image matching according to the image feature vector and the skin disease image database to obtain a similar image and corresponding text information; The question library question module is used for determining text feature vectors of the question library by using the question library according to the text information corresponding to the similar images; And the skin disease determining module is used for determining skin disease types and probabilities by using a skin disease classification model according to the text feature vector of the question bank and the text feature vector. Optionally, the similar image matching module specifically includes: A calculating unit for calculating cosine similarity distance between the image feature vector and the image feature vector in the skin disease image database; and the matching unit is used for determining similar images and corresponding text information according to the cosine similar distance. Optionally, the question library questioning module specifically includes: the questioning unit is used for questioning by utilizing questions in a question bank according to the text information corresponding to the similar images