An Automated System for the Early Detection of Dysgraphia using Deep Learning Algorithms
C Sharmila, N. Shanthi, S. Santhiya, E Saran, K Sri Rakesh, R Sruthi
Abstract
Children with specific disability challenges and impede their academic development in today’s world. While some of the disorders are clearly evident to the untrained eye, others are difficult to locate and require special care. Dysgraphia is one such ailment. Dysgraphia is a learning disorder of written expression that impairs coherence and handwriting in addition to overall writing abilities. It is both a transcribing disability and a Special Learning Disability (SLD), which means that it is a writing issue accompanied by poor handwriting. Learning disabilities have an immediate impact on the brain. It is highly expensive and taxing for youngsters to go through the routine procedures used to diagnose this disease. However, they are not entirely standardised for evaluation. Numerous sophisticated computational approaches have been proposed, bearing a wide variety of executions. Improvements in Deep Learning methods have shown to be useful in automating this identification procedure. An automated system is developed for the screening of learning disability. Deep Learning techniques are used in this work to create the system for detecting learning disabilities. The model evaluates the handwriting of child to determine if the child has a disorder or not. Convolutional Neural Network (CNN), Inception V3, RESNET, and VGG16 are employed and the disability detection results are examined.