DYSIGN: Towards Computational Screening of Dyslexia and Dysgraphia Based on Handwriting Quality

Specific Learning Difficulties such as Dyslexia and Dysgraphia are characterized by struggles in reading and writing. Their diagnosis and intervention are critical as if left unattended, they can cause hindrance in academic activity, self-esteem, and long-term quality of life. Owing to the complex traditional processes for diagnosis, social stigma, and the general lack of availability of remedial therapists and clinical psychologists in Pakistan, this study explores the potential of handwriting quality features to be used in computationally screening for SLDs to make screening more accessible. This project consists of exploratory data analysis of handwriting scans of 25 children thus far, in the age group of 5 to 15, generating various handwriting quality features and using classification models to assess their potential. Our preliminary results are promising, with approximately 80% accuracy, thus showing potential for increased accuracy when paired with larger data samples and further feature generation.