Handwriting Analysis for Dysgraphia Using Machine Learning

Learning disabilities specifically Dyslexia, Dysgraphia and Dyscalculia affect around 10% of children, holding back with their academic performance along with long term consequences such as taking a toll on their self-esteem and even their attitude towards education. As a neurological condition, the poor motor skills and motor coordination, these signs are reflected upon multiple aspects of written skills. An early diagnosis enables quick detection and have remedies to overcome writing-related challenges and reach their full potential. In this paper we have shown identification of dysgraphia by understanding deep leaning algorithms that have been already done by researchers (Mekyska et al. in IEEE Trans Hum Mach Syst 47:235–248, 2019) Performance of dysgraphia students’ handwriting with different models is discussed and implementation of CNN and RNN on dataset found on Kaggle have been accomplished. This study’s objectives were to analyse and discriminate the sample of handwriting capabilities of students suffering with dysgraphia handwriting. With the help of deep learning techniques, we have applied CNN and RNN on existing data and find out result on training dataset.