Analysis of the CNN Models Performance to Detect Handwriting Difficulties
Nisha Ameya Vanjari, Prasanna J. Shete · 2024
Handwriting difficulties may be the start of specific learning disability like dysgraphia. Dysgraphia is a disability that impairs a person’s ability to communicate symbols and words in writing. It has a detrimental effect on students’ academic performance and general well-being. Early intervention for those in need can be facilitated by expanding the availability of dysgraphia testing to a wider audience through the use of automated processes. Raising awareness of the issue of dysgraphia and its impact on society is another goal of this paper. In order to detect handwriting impaired by dysgraphia, we used a deep learning approach in this research. We assembled a dataset of handwritten to accomplish this goal. To determine whether handwriting is impacted by dysgraphia, we have used a deep learning algorithm like VGG16, ResNet50 and CNN1, and we get accuracies 72%, 62% and 84% respectively from pre-trained models. All are based on feed forward approach.