CN-121983275-A - Deep learning-based oral disease identification system

CN121983275ACN 121983275 ACN121983275 ACN 121983275ACN-121983275-A

Abstract

The invention relates to the technical field of oral diseases and discloses an oral disease identification system based on deep learning, which comprises an image acquisition module, an oral disease identification module, a data acquisition module and an analysis processing module, wherein the image acquisition module is used for acquiring an oral image of a target user, the oral disease identification module is used for receiving the oral image and carrying out oral disease identification on the oral image to obtain a corresponding disease category, the data acquisition module is used for periodically acquiring oral data of the target user, the analysis processing module is used for receiving the oral data and adjusting the acquisition period of the oral data according to the oral data, comparing the acquisition period with a preset period and adjusting an oral monitoring strategy according to the comparison result. According to the deep learning-based oral disease identification system, a target user is helped to know the oral health development change of the user in real time, so that effective protection and control are carried out in advance, and the oral disease monitoring efficiency is improved.

Inventors

ZHANG RUI
WANG SHUYUE
SHAO WENZHU
LIU YAHUI
LIU CHANGZHENG
FENG ZHIQIANG
ZHOU ZHENG
MA XUEHUA
GAO YAHONG
ZHANG WENXING

Assignees

石河子大学

Dates

Publication Date: 20260505
Application Date: 20240308

Claims (8)

1. An oral disease identification system based on deep learning, comprising: the image acquisition module is used for acquiring an oral cavity image of a target user; the oral disease identification model is used for receiving the oral image and identifying the oral disease of the oral image to obtain a corresponding disease category; The data acquisition module is used for periodically acquiring the oral cavity data of the target user; The analysis processing module is used for receiving the oral data, adjusting the acquisition period of the oral data according to the oral data, comparing the acquisition period with a preset period and adjusting an oral monitoring strategy according to the comparison result; wherein the oral disease identification model is a trained machine learning model.
2. The deep learning based oral disease identification system of claim 1 wherein the process of adjusting the period of collection of the oral data based on the oral data comprises: counting the time-varying curves of all oral parameters according to the oral data; Obtaining a time-varying curve of each oral parameter variable according to the oral data; by the formula: Calculating an adjustment coefficient gamma t of the acquisition period; Calculating an adjusted acquisition period according to the adjustment coefficient gamma t ; Wherein n is the number of items of the oral parameters, C i is the time-varying curve of the ith oral parameter, C ilow ,C iup is a standard interval corresponding to the ith oral parameter, deltaC i0 is a preset difference value, C i 'is the time-varying curve of the variation of the ith oral parameter, C ith ' is the variation threshold of the ith oral parameter, alpha i is the weight coefficient of the ith oral parameter, beta 1 ,β 2 is a preset weight coefficient, and y is a conversion function.
3. The deep learning based oral disease identification system of claim 2 wherein the process of calculating an adjusted acquisition period from the adjustment factor γ t comprises: by the formula: T t ＝T 0 +γ t *Δt calculating an adjusted acquisition period T t ; Wherein T 0 is the initial acquisition period, and Δt is the preset adjustment interval.
4. The deep learning based oral disease identification system of claim 3 wherein the initial acquisition cycle acquisition process comprises: obtaining a corresponding disease state corresponding interval according to the disease category; and obtaining an initial acquisition period according to the disease state corresponding interval.
5. The deep learning based oral disease identification system of claim 3, wherein the process of comparing the collection period to a predetermined period and adjusting the oral monitoring strategy based on the comparison result comprises: Comparing the adjusted acquisition period T t with a preset period T min : if T t >T min , taking the T t as an acquisition period to acquire the next period, and recording the change of the adjustment coefficient in real time; Otherwise, an oral disease early warning signal is sent out, the preset period T min is used as an acquisition period to acquire the next period, and the current oral image and the oral data record of the whole development period are stored.
6. The deep learning based oral disease identification system of claim 5 wherein the analysis processing module further comprises determining oral hygiene habit status of the target user based on a law of change of the adjustment factor: by the formula: calculating a rule coefficient rho; judging the oral hygiene habit state of the target user according to the rule coefficient rho; Wherein, gamma t ,γ t+1 ,γ t+2 ,…,γ t+m-2 ,γ t+m-1 is m groups of acquisition periods larger than the preset period.
7. The deep learning based oral disease identification system of claim 6 wherein the process of determining the oral hygiene habit status of the target user based on the law coefficient ρ comprises: when the acquisition period gradually becomes smaller, judging Whether a preset rule range of reduction is satisfied: if the signal meets the requirement, sending out an oral disease early warning signal; Otherwise, sending out an oral hygiene habit adjustment signal; When the acquisition period gradually becomes larger, judging Whether a preset rule range of enlargement is satisfied: If so, sending out an oral hygiene habit maintaining signal; Otherwise, sending out an oral hygiene habit adjustment signal; Wherein σ 1 and σ 2 are preset rule coefficients.
8. The deep learning based oral disease identification system of claim 5 further comprising a weight updating module for periodic training of the oral disease identification model based on an expanded sample set, updating network weights; the extended sample set is a set of oral cavity images stored when the collection period is not greater than the preset period.

Description

Deep learning-based oral disease identification system Technical Field The invention relates to the technical field of oral diseases, in particular to an oral disease identification system based on deep learning. Background Oral health is an important component of general health, and is one of the ten major criteria for human health. Oral diseases are common diseases and frequently-occurring diseases affecting the health of residents, and have high prevalence rate and wide scope of spread. Oral diseases affect oral health, toothache, tooth loss, poor occlusion, etc. can cause reduced or lost chewing function, and inflammation of gingiva, tooth root or dental nerve, etc. can cause pain, resulting in organism discomfort. Periodontal disease has a close relationship with general health and may trigger, induce or exacerbate systemic disease. Chronic inflammatory diseases such as periodontitis can cause bacteria to enter blood and moving cells, produce inflammatory biological signals that can damage the body, and can cause type II diabetes. Periodontal disease bacteria can infect cardiovascular tissues through blood, and once they enter the blood stream, oral bacteria adhere to fatty acids in the coronary arteries, leading to the formation of blood clots, and thus easily induce heart disease. However, at present, the knowledge rate of oral knowledge is still low, the knowledge and understanding of oral diseases are mainly from clinical oral examination, most people seek to seek medical attention only after obvious pain symptoms appear, and the oral health examination and the knowledge of oral health conditions are lacking, so that oral diseases are not effectively prevented and controlled, and therefore, an oral disease identification system based on deep learning is provided for solving the problems. Disclosure of Invention The invention aims to provide an oral disease identification system based on deep learning, which solves the following technical problems: How to provide an oral disease recognition system capable of timely monitoring oral problems. The aim of the invention can be achieved by the following technical scheme: An oral disease identification system based on deep learning, comprising: the image acquisition module is used for acquiring an oral cavity image of a target user; the oral disease identification model is used for receiving the oral image and identifying the oral disease of the oral image to obtain a corresponding disease category; The data acquisition module is used for periodically acquiring the oral cavity data of the target user; The analysis processing module is used for receiving the oral data, adjusting the acquisition period of the oral data according to the oral data, comparing the acquisition period with a preset period and adjusting an oral monitoring strategy according to the comparison result; wherein the oral disease identification model is a trained machine learning model. Preferably, the process of adjusting the collection period of the oral data according to the oral data comprises: counting the time-varying curves of all oral parameters according to the oral data; Obtaining a time-varying curve of each oral parameter variable according to the oral data; by the formula: Calculating an adjustment coefficient gamma t of the acquisition period; Calculating an adjusted acquisition period according to the adjustment coefficient gamma t; Wherein n is the number of items of the oral parameters, C i is the time-varying curve of the ith oral parameter, C ilow,Ciup is a standard interval corresponding to the ith oral parameter, deltaC i0 is a preset difference value, C i 'is the time-varying curve of the variation of the ith oral parameter, C ith' is the variation threshold of the ith oral parameter, alpha i is the weight coefficient of the ith oral parameter, beta 1,β2 is a preset weight coefficient, and y is a conversion function. Preferably, the process of calculating the adjusted acquisition period according to the adjustment coefficient gamma t includes: by the formula: Tt＝T0+γt*Δt calculating an adjusted acquisition period T t; Wherein T 0 is the initial acquisition period, and Δt is the preset adjustment interval. Preferably, the acquiring process of the initial acquisition period includes: obtaining a corresponding disease state corresponding interval according to the disease category; and obtaining an initial acquisition period according to the disease state corresponding interval. Preferably, the process of comparing the collection period with a preset period and adjusting the oral monitoring strategy according to the comparison result includes: Comparing the adjusted acquisition period T t with a preset period T min: if T t>Tmin, taking the T t as an acquisition period to acquire the next period, and recording the change of the adjustment coefficient in real time; Otherwise, an oral disease early warning signal is sent out, the preset period T min is used as an acquisiti