Technologies Of Intelligent Diagnosis And Rehabilitation Training For Children With ASD

Autism spectrum disorder(ASD)is a neurodevelopmental disorder that seriously affects cognition and social behavior.Childhood is the key to rehabilitation.Early diagnosis and effective intervention are crucial for the rehabilitation of ASD patients.Computer-aided diagnosis and rehabilitation training technology have great potential to improve the efficiency of diagnosis and rehabilitation training.The research on effective computer-aided diagnosis technology and rehabilitation training theory and method have important theoretical significance and application value.Computer-aided diagnosis and training for ASD children are current research hotspots.Some exploratory research work has been combined with virtual reality(VR)and mixed reality(MR)technologies.These research results can help ASD children improve their abilities in perception and cognition,emotional understanding,and social and life skills.However,the current training model still needs to be improved in many aspects,especially for young ASD children.There is little involvement in the auxiliary diagnosis technology based on externalizing behaviors such as stereotyped behavior and attention execution.Therefore,the diagnosis and rehabilitation training methods in line with the individual cognitive and behavioral characteristics of ASD need to be further studied to improve the efficiency of ASD diagnosis and rehabilitation training.This study aims to explore the new model of rehabilitation training for ASD children based on the cognitive and behavioral characteristics of ASD children,break through the key technologies of intelligent diagnosis and rehabilitation training,and contribute to improving the diagnosis level and rehabilitation training effect of ASD children.Combined with the cognitive characteristics of ASD children,this paper focuses on four aspects of research,including a mixed reality-based rehabilitation training model,the interactive accuracy of MR systems,computer-aided diagnostic techniques for ASD children,and the evaluation of training effects.Major innovation and academic contributions include:Firstly,a new method of MR rehabilitation training for ASD children is proposed.By transforming the aquarium in life,an MR interactive aquarium training mode including MR course,MR display and interaction,and computer-aided diagnosis technology is presented,which can be effectively used in the rehabilitation training of ASD children.Secondly,an interactive technology to improve touch accuracy is proposed.Aiming at the tracking and touch errors caused by water surface refraction in MR aquarium,a viewpoint interaction optimization method combining water refraction and eye-tracking pre-calibration is proposed,which improves users’ adaptability and touch interaction accuracy.Combined with the three Gaussian models,a new method to improve the accuracy of touch interaction of moving targets is proposed.Through the research of the above contents,this thesis draws:Thirdly,a video-based diagnosis method of stereotyped movement in ASD children is proposed.The data set of stereotyped movement of ASD children based on video was established;The TextCNN+BiLSTM+Attention model is proposed.After training the stereotyped movement data set of ASD children,a good auxiliary diagnosis effect of stereotyped movement is realized.The accuracy of diagnosis of five types of stereotyped movement is 80.33%.Fourthly,a diagnosis method of attention execution(executive control)of ASD children based on touch behavior modeling is proposed.The three Gauss model is used to model the touchpoint data set of ASD children in moving target selection,and the touchpoint distribution model of ASD children is obtained.By comparing the click drop point distribution with that of non-ASD children,it is found that ASD children are not good at perceiving and predicting the location of moving targets.They have the characteristics of scattered touch drop points and large overall uncertainty,providing doctors with an auxiliary diagnosis basis for paying attention to execution problems.Finally,this paper evaluates the effect of the proposed new method of MR rehabilitation training in clinical rehabilitation.Empirical research shows that this model can effectively improve the feedback,attention,interest,and cognitive ability of ASD children in training and show apparent training gain.