Thu Nov 21 13:33:25 UTC 2024: ## Bengaluru Researchers Develop AI System for Early Autism Detection

**Bengaluru, November 21, 2024** – Researchers at the International Institute of Information Technology Bangalore (IIITB) have developed an AI-powered system that could revolutionize early detection of Autism Spectrum Disorder (ASD). The system, showcased at the Bengaluru Tech Summit, uses computer vision to analyze children’s behavior, achieving an 82% accuracy rate in predicting ASD in children aged 18 to 42 months.

The study, a collaboration between IIITB and St. John’s Hospital’s Centre for Advanced Research and Excellence in Autism and Developmental Disorders, employs a novel play protocol specifically designed for the Indian context. This protocol involves structured and unstructured play activities recorded on video. Deep learning techniques are then used to analyze 26 behavioral markers across facial expressions, social communication, and play behavior.

“Early detection of ASD is crucial, and our approach addresses unique challenges within the Indian context,” explained Dinesh Babu Jayagopi, HoD of Data Science and Artificial Intelligence at IIITB. The system analyzes video recordings for patterns like eye contact and repetitive actions, eliminating the need for manual intervention. Privacy protections are built into the system.

The research utilized data from 100 children diagnosed with ASD and 40 without. While the current accuracy is promising, further large-scale testing is needed to validate the system’s effectiveness. The researchers aim to simplify the system for use by community health workers, such as ASHA workers, even in remote areas. The project is funded by the Machine INtelligence and RObotics (MINRO) Center at IIITB.

Read More