Novel framework for dyslexia diagnosis in children in Al Kharj region with super-resolution generative adversarial network and transfer learning technique

Abstract Learning disabilities like dyslexia are commonly prevalent among young school children. Dyslexia is a neurological disorder that can drastically impact a child’s academic life and mental health, often resulting in low self-esteem. This research study aims to design and implement an easy-to-use computer-aided diagnosis tool for the early detection of dyslexia, ensuring that dyslexic children can receive timely support from teachers and experts. The novel framework, which incorporates Super-Resolution Generative Adversarial Network, and a custom-built convolutional neural network model based on transfer learning technique, achieves 92.52% accuracy in the classification of handwriting of either dyslexic or non-dyslexic individuals.