Abstract—Analysis of the behavior of children is important for the early detection of developmental disorders such as Autism Spectrum Disorder (ASD) that is usually characterized by impairments in social and communication skills and repetitive and stereotyped behaviors as several studies about these behaviors revealed indicative of ASD in recorded videos of children later diagnosed with ASD. Developing a computational approach that is standardized and objective for the behavioral phenotyping of children with autism spectral disorder is significant for data gathering for creating prediction model for screening and monitoring of these patterns. We create a scalable, baseline application using computer vision algorithms and methodologies to capture four simple biomarkers data reliably for visual attention tracking using head pose estimation, specific observable behavioral patterns measuring blink rate and body posture, and morphological anomalies examining open mouth appearance from children with ASD. We conclude that it is feasible to quantitatively measure these behavioral phenotypes and there are promising results from measuring simple biomarkers showing distinguishing results for at least three biomarkers.
Index Terms—computer vision, autism spectrum disorder, behavioral phenotyping, pattern analysis
Cite: James-Andrew R. Sarmiento and Prospero Naval Jr, "Behavioral Phenotyping for Autism Spectrum Disorder Biomarkers Using Computer Vision," Journal of Image and Graphics, Vol. 8, No. 2, pp. 47-52, June 2020. doi: 10.18178/joig.8.2.47-52
Copyright © 2020 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.
Copyright © 2012-2020 Journal of Image and Graphics, All Rights Reserved