Development of Intelligent Audio-Based Support Systems for Smart User Assistance

Sensory processing disorders in individuals with autism can result in increased sensitivity to auditory stimuli. Sonic-aid aims to address this challenge by developing an AI- powered tool capable of classifying audio input as speech or music. Using Sub-band Coding (SBC) for feature extraction and employing Support Vector Machines (SVM) and Gaussian Mix- ture Models (GMM) for classification, Sonic-aid enables real-time analysis and therapeutic sound filtering. The system is optimized for scalability, visualization, and user-friendly deployment using Fast-API and TensorFlow. This paper outlines the methodology, implementation and realworld applications of the Sonic-Aid system in therapeutic and accessibility contexts