A Machine Learning Framework for Early Dysgraphia Screening and Support Through Handwriting

Dysgraphia is a learning difficulty that significantly affects handwriting, spelling, and fine motor control, often leading to academic struggles and reduced confidence in children. Early identification is essential, yet current diagnostic practices are time- consuming, costly, and dependent on specialist availability, leaving many cases undetected, especially in under-resourced settings. This paper proposes a technology-assisted framework for early screening and support of dysgraphia using machine learning– driven handwriting analysis combined with real-time classroom observations. The system leverages mobile-based tools employing techniques such as edge detection and convolutional neural networks (CNNs) to analyse handwriting patterns, while integrating insights from teachers and parents for a holistic evaluation. Beyond detection, the framework provides actionable, real- time feedback to support educators, families, and clinicians in developing personalized learning strategies, classroom adjustments, and ongoing progress tracking. By making screening accessible, cost- effective, and scalable, the approach reduces reliance on clinical resources and promotes inclusive learning environments. Ultimately, it empowers children with dysgraphia to receive timely interventions, fostering both academic achievement and social confidence from the earliest stages of education.