Enhancing Emotional Literacy in Children’s Writing: An AI-Driven Assistant for Emotion Detection and Visual Reinforcement

Abstract

This project talks about how to make and test an AI-powered writing assistant that is supposed to help kids write better by teaching them how to understand their feelings. The system uses Stable Diffusion to make pictures that fit the story and a carefully tuned RoBERTa model to find feelings. It tells you right away how sad the stories for kids are. It can tell when a child is angry, scared, sad, or happy, which helps them write about and understand how they feel. Flask is used for the back end and react is used for the front end of the web app. It allows emotion detection and visual feedback to work together without any problems. The AI model learned from the PoKi dataset to better understand how to recognize target emotions. Stable Diffusion, on the other hand, makes pictures to help people understand how they feel. The study looks at how well the system can find emotions by looking at things like precision, recall, and F1-score, as well as how interested users are. The results show that the system works well to help kids write about how they feel in a clearer way. The project gives teachers a tool that many students can use to help them write better and understand their feelings better. This is one more way that AI is used in schools.