Software Engineering Department Seminars
Title: A customized CNN for Multi class Autism’s classification using the CARS ADK dataset .
Presented by: M. Rana Laith
On: 28 April 2026
Abstract: customized CNN for multiclass autism classification using the Childhood Autism Rating Scale – Autism Dataset for Kids (CARS-ADK) involves designing a specialized, lightweight architecture to process behavioral features, usually extracted from clinical questionnaires and behavioral observations, rather than images. Based on the standard interpretation of CARS scores, the model is typically designed for three-class classification: Non-Autism (<30), Mild/Moderate Autism (30-36.5), and Severe Autism.
