Related Experiment Video
Updated: Jul 6, 2025
09:16
Use of a Video Scoring Anchor for Rapid Serial Assessment of Social Communication in Toddlers
Published on: March 14, 2018
10.3K
Seeing through a robot's eyes: A cross-sectional exploratory study in developing a robotic screening technology for
Wing-Chee So1, Elsa Wong2, Wingo Ng2
1Department of Educational Psychology, The Chinese University of Hong Kong, Shatin, Hong Kong.
Summary
This study shows social robots can screen for autism by tracking children's eye gaze. The HUMANE robot demonstrated high accuracy in detecting atypical eye patterns in children aged 3-8.
Area of Science:
- Robotics
- Developmental Psychology
- Artificial Intelligence
Background:
- Autism Spectrum Disorder (ASD) screening tools require development for scalability and reliability.
- Social robots offer novel technological approaches for behavioral assessment.
Purpose of the Study:
- To develop and validate a social robot-based screening tool for autism.
- To assess the efficacy of the HUMANE robot in detecting eye gaze patterns indicative of autism.
Main Methods:
- An exploratory, cross-sectional case-control study involving 199 children aged 3-8.
- Utilized the HUMANE social robot with computer vision to detect eye gaze direction.
- Compared robot's eye gaze detection with human raters and the Autism Spectrum Quotient (ASQ) as a reference standard.
Main Results:
- Excellent interrater reliability (0.90) between human raters and the HUMANE robot.
- High sensitivity (≥0.88) and specificity (≥0.88) for the robot's screening index tests.
- Diagnostic Odds Ratios exceeding 190, indicating strong predictive value.
Conclusions:
- Social robots can reliably detect atypical eye gaze patterns in children.
- The HUMANE robot shows significant potential as a scalable tool for autism screening.
- Future research can explore integrating social robots into early autism detection pathways.
