Predicting ASD diagnosis in children with synthetic and image-based eye gaze data
作者:
Highlights:
• Atypical visual attention associated with autism spectrum disorders can be studied through eye gaze behavior of children at a very young age.
• Autistic gaze viewing patterns in a free-viewing paradigm studied using computer vision.
• Correlation between scan-path fixations and high-level image semantics extracted through deep learning result in ASD prediction accuracy up to 62%
• Study finds significant class differences by applying machine learning to raw scan-path fixations and duration; verification on larger dataset needed.
摘要
•Atypical visual attention associated with autism spectrum disorders can be studied through eye gaze behavior of children at a very young age.•Autistic gaze viewing patterns in a free-viewing paradigm studied using computer vision.•Correlation between scan-path fixations and high-level image semantics extracted through deep learning result in ASD prediction accuracy up to 62%•Study finds significant class differences by applying machine learning to raw scan-path fixations and duration; verification on larger dataset needed.
论文关键词:Autism spectrum disorders,Eye gaze data,Deep learning
论文评审过程:Received 29 May 2020, Revised 28 January 2021, Accepted 1 February 2021, Available online 16 February 2021, Version of Record 3 March 2021.
论文官网地址:https://doi.org/10.1016/j.image.2021.116198