Research On Hand,Foot And Mouth Disease Assisted Detection Technology Based On Deep Learning And Machine Vision

Hand,foot and mouth disease is a high recessive infectious disease that occurs in children’s hands,feet,mouth,and other parts,and has the risk of complications and even death.Early screening and detection reduce the infection and mortality of hand,foot and mouth disease Effective means.At present,HFMD detection and recognition based on artificial intelligence is implemented at home and abroad.Most of its research focuses on the use of machine learning algorithms to classify clinical case data such as virus types,peak temperatures,etc.Use this to identify hand,foot and mouth disease.In recent years,more and more researchers have applied deep learning to medical image target recognition and detection,and often have reached doctor-level accuracy in a large number of diagnostic tasks.If the skin case characteristics of the hands,mouth,and other parts can be used to screen and detect early hand-foot-mouth disease,the condition can be found in time to prevent further development and spread of the condition.Based on this,this paper proposes a method for assisted detection of hand,foot and mouth disease based on machine vision and deep learning.The main research contents include:(1)By referring to the clinical diagnosis standard of hand,foot and mouth disease,a visual algorithm is proposed to detect the characteristics of hand,foot and mouth disease skin cases in the image,so as to realize the auxiliary detection and screening of hand,foot and mouth disease.And build a data set based on the different symptoms of hand,foot and mouth disease,and expand the data sample through data augmentation technology.(2)Perform experimental tests with a single-stage detection algorithm and a two-stage detection algorithm.based on the detection results of different algorithms,the shortcomings of the single-stage detection algorithm applied to the research of this paper and the feasibility of the two-stage detection algorithm applied to the detection task of this paper are analyzed.(3)Improve the original detection algorithm by optimizing the basic feature extraction network and increasing the CBAM attention mechanism.The experimental results show that the improved detection algorithm in this paper has a good detection effect on the characteristics of hand,foot,and mouth disease skin cases,and the accuracy rate is improved by about 3% compared with the original detection algorithm.(4)In view of the problem that the original detection algorithm has a high false detection rate,the original detection mechanism is improved by adding a method of judging logic to exclude non-real targets.Including non-real targets in non-skin areas based on YCrCb + Otsu skin color detection,and non-real targets in non-oral areas based on HOG + SVM oral detection and ERT keypoint positioning.Experimental results show that the improved detection mechanism in this paper can effectively reduce the false detection rate of the original detection algorithm.