Abstract
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder characterized by impairments in social communication and interaction, as well as restricted and repetitive behaviors. ASD affects individuals across various aspects of their lives, including their cognitive, social, and motor skills. Early detection and intervention are crucial for improving outcomes and providing appropriate support to individuals with ASD. Handwriting is an essential skill that children develop early in life, and it provides valuable insights into their motor control, coordination, and cognitive processes. Good handwriting is crucial for academic progress, social and communicative development, and development of self-esteem. By identifying the handwriting characteristics exhibited by ASD children, we aim to develop a reliable and objective method for detecting ASD. We aim to identify the handwriting characteristics associated with ASD using deep learning techniques. The characteristics of Chinese characters, such as straight lines and turns, are considered in analyzing the writing characteristics of ASD children. We present a dataset of ASD children's handwriting in Traditional Chinese and fine-tunes three models to build a classification model. Experiment results demonstrate the model's effectiveness in distinguishing between the handwriting of ASD children and typically developing children, achieving F1 score of 93.6%. The study provides valuable insights into handwriting characteristics in ASD and suggests directions for future research in this area.