Identifying Handwriting Difficulties in Children in Devanagari Script Using Machine Learning

Dysgraphia is a medical disorder in children, due to which children face difficulties in writing. Children suffering from dysgraphia find it difficult in holding a pencil, alignment, differentiating the strokes, curves, and sizing of the alphabets, etc. It is more challenging in scripts where there are more alphabets, curves, and strokes. In India, about 75% of school-going kids write in Devanagari script (includes Hindi, Marathi, Gujarati) as a medium of study or second language. Diagnosis of writing difficulties in the Devanagari script can help the child to change the language or get assistance. Many researchers have come up with diagnosing dysgraphia in languages like English, Hebrew, etc. The current study aims at developing a model to identify children with dysgraphia and without dysgraphia in Devanagari script based on their performance characteristics as per the BHK algorithm. 52 students handwriting samples were collected, and feature analysis is done based on BHK and machine learning model using the KNN algorithm for the predictions.