MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641108583 A) filed by Kcg College Of Technology on September 10, 2026, for A Data-Driven Approach To Early-Stage Autism Spectrum Disorder Detection.
Inventors include Dr. Adline Freeda; Chris; Chris Livingstone; and Gopikrishnan A.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: ABSTRACT Early identification of autism spectrum disorder (ASD) plays a critical role in enabling the implementation of prompt intervention and enhancing overall developmental outcomes. While clinical assessment is conventionally used for diagnoses, traditional screening approaches are typically costly and time-consuming, often limiting early administration. This work introduces a multimodal screening system that combines questionnaire data with automated facial expression recognition for early detection of the risks of ASD. The developed system first leverages structured questionnaires, extracts behaviors associated with social communication and restricted, repetitive behavior, and quantities them. Then, the expressions of children exposed to emotional visual stimuli are recorded and deep neural network-based models extract discriminative emotional features (such as frequency, latency and dominance of neutrality), from the facial expressions. This information from behavioral and visual streams are then combined using a machine learning classifier for the prediction of the probability of ASD risk. Experimental results demonstrate the system achieves enhanced accuracy and reliability by multimodal data integration.
Disclaimer: Curated by HT Syndication.
