Assist Dyslexic Children in Improving Readability Using Advanced Machine Learning Techniques

Abstract

Lorem Dyslexia is a learning disorder common in children, characterized by difficulties with reading and language processing skills. Children with Dyslexia often need more time to process these skills than their peers. Early intervention can help mitigate these challenges. Dyslexia, often identified in children at an early age, poses challenges in decoding letters, recognizing sounds, understanding alphabets, and related skills. This study concentrates on enhancing the reading proficiency of dyslexic children, specifically in the Punjabi language. The research targets children above six years old who encounter difficulties in reading and exploring various hardware and software techniques to address these issues. The study entails recording 13,000 audio samples containing both accurate and inaccurate pronunciations of two and three-letter Punjabi words. These audio samples serve as the primary input for the system. To achieve this goal, a dataset comprising Punjabi words pronounced by children of diverse age groups through visits to different schools and then system trained using this dataset. Children will be presented with a set of words to read. If they pronounce a word incorrectly, the system will provide the correct speech, aiding in the training of the children. In essence, this paper presents a prototype model to support dyslexic children in enhancing their reading skills, with a specific focus on the Punjabi language.