Development of potential dysgraphia handwriting dataset

Area of Science:

• Educational Psychology
• Developmental Neuroscience
• Computational Linguistics

Background:

• Dysgraphia, a learning disability affecting writing, presents challenges in academic settings.
• Early identification and intervention are crucial for supporting children with dysgraphia.
• A lack of specialized datasets can hinder research and development of diagnostic tools.

Purpose of the Study:

• To present a novel dataset of offline handwriting samples from Malaysian schoolchildren.
• To facilitate research on identifying potential dysgraphia in young learners.
• To provide a resource for developing and validating automated handwriting analysis tools.

Main Methods:

• Collected handwriting samples of Malay sentences from primary school students.
• Included students undergoing intervention by the Malaysia Dyslexia Association (PDM).
• Digitalized scanned handwriting, pre-processed images using binary conversion and color inversion.
• Classified images into 'potential dysgraphia' and 'low potential dysgraphia' categories.

Main Results:

• The dataset contains 249 black and white handwriting images.
• Data was collected from 83 participants.
• 114 images were classified as potential dysgraphia, and 135 as low potential dysgraphia.

Conclusions:

• The dataset offers a valuable resource for studying dysgraphia in a specific linguistic and cultural context.
• This data can advance the development of assistive technologies for children with writing difficulties.
• Further research can utilize this dataset to explore the characteristics of handwriting in children with and without dysgraphia.