An AI-based Intelligent Social Companion app is developed based on Kivy, a Cross-Platform Python Framework. The app aimed to cultivate better social skills and to improve the social behaviour of persons with Intellectual Disabilities (PWIDs). The developed solution has 3 major features:
[Conversation Simulator] The conversation simulator aims to cultivate socially acceptable behaviours within PWIDs through remote means. Provides a conversation prompt and grades the response according to how appropriate it is.
[Sign Language Neural Network] A sign language recognition system is developed using a Multilayer Perceptron (MLP) Deep Learning model and optimised with oneAPI Deep Neural Network Library (oneDNN).
[Anti Spam Engine] An anti-spam engine is developed using OpenCV Computer Vision libraries to perform real-time template matching across different screens, which accurately detects and prevents spamming.
This provides a safe space for people with IDs to learn socially acceptable ways to communicate and express their thoughts and feelings without offending others.
Developed by: py.jamas (Kiran, Kazu and Yanzhao (not in photo))