Virtual Booth

Dyslexia affects reading fluency, decoding, and comprehension, making standardized instructional materials (i.e., textbooks) challenging in inclusive classrooms. This project introduces a personalized text simplification model backboned by a hybrid deep learning framework. In the first phase, student classification is performed using handwriting images through a hybrid framework combining Convolutional Neural Networks and Vision Transformers, assigining students into mild, moderate, or severe reading levels. Sentence complexity is then analyzed using shallow linguistic features and deep neural networks. Finally, an explicit editing algorithm restructures and simplifies text. This innovation supports differentiated instruction, enhances reading comprehension, and aligns with Sustainable Development Goal 4 on equitable and inclusive education.