摘要
- OBJECTIVE: Autism spectrum disorder is a group of neurodevelopmental disor-ders. The disease's etiology is unclear, and there is no specific drug treatment for the core symp-toms of autism spectrum disease. The study aims to explore effective intervention methods for children with autism spectrum disorders. MATERIALS AND METHODS: This paper proposes a visual strategy intervention meth-od for children with autism spectrum disorders. This method combines feature extraction and abnormal behavior detection and can use a vi-sual cue strategy to integrate children into so-cial groups. Firstly, the spatial-temporal feature fusion structure is added to extract behavioral features from children, and the spatial informa-tion contained in MotionNet is fused with tem-poral features. Optical Flow Feature (OFF) sub-network is added to the optical flow extraction feature network. Each layer feature is input to the OFF subnet to extract the time feature fur-ther. Then, a behavior detection method based on the sequential pool is proposed. This method combines attention mechanism and clustering pool to effectively describe human behavior dy-namics in the long, redundant video under com-plex background. Finally, feature extraction and behavior detection experiments are carried out on SDUFall, Weizmann, and HMDB51 data sets. RESULTS: The model's accuracy is still slight-ly higher than others in that only the video Red -Green-Blue (RGB) frame is used as input. Com-pared with OFF, SDUFall can reach 88.64%, and HMDB51 can only reach 63.81%. In contrast, the proposed model can reach 72.09%, higher than others. The descriptor obtained the best result of 92.57%, which is 3.64%, 2.58%, and 1.73% higher than the other three comparison descrip-tors. The data show that the method presented here is effective and has advantages in detect-ing children's abnormal behavior. CONCLUSIONS: This method and visual inter-vention for children with autism spectrum disor-ders can help them to overcome social barriers.
学科与关键词
- Automatic classification of ASD children using appearance-based features from videosLi, Jing*; Chen, Zejin; Li, Gongfa; Ouyang, Gaoxiang; Li, XiaoliNEUROCOMPUTING, 2022, 470: 40-50.被引 2
- Identifying Visual Attention Features Accurately Discerning Between Autism and Typically Developing: a Deep Learning FrameworkXie, Jin; Wang, Longfei; Webster, Paula; Yao, Yang; Sun, Jiayao; Wang, Shuo*; Zhou, Huihui*Interdisciplinary Sciences-Computational Life Sciences, 2022, 14(3): 639-651.被引 6
- Application of Machine Learning Techniques to Detect the Children with Autism Spectrum DisorderLiao, Mengyi*; Duan, Hengyao*; Wang, GuangshuaiJournal of Healthcare Engineering, 2022, 2022: 9340027.被引 15
