Early Autism Detection Mobile Application

I worked on an autism detection mobile application for children in a joint project with the Convergent Technologies Research Center (NBIC) under Prof. Y. Yaghoobzadeh’s supervision.

The app aims to diagnose autism at an earlier age than traditional methods by analyzing children’s behavior, speech, and facial expressions. My responsibility in this project was to integrate a voice assistant to the app which gives instructions to the user and processes feedback from the child.

I fine-tuned the SeamlessM4T model and utilized the ChatGPT API to handle verbal inputs, convert them to text, process the data, and generate verbal responses. This implementation achieved satisfactory performance, significantly enhancing the app’s ability to interact with users effectively.