ABSTRACT Dysgraphia, a key component of specific learning disorders, can be computationally identified using artificial intelligence (AI) models based on offline handwriting features. This study addresses the challenges of processing large volumes of inconsistent handwriting data by introducing H2FCD, a novel classification model that fuses deep feature extraction techniques for offline handwriting analysis. The H2FCD model combines feature vectors derived from resized handwriting images and histogram‐based extractions using a convolutional neural network (CNN) with a ResNet‐18 architecture and proposes a new composite of deep feature extractor layer structure. Dysgraphia severity levels were classified using three machine learning models, with the support vector machine (SVM) achieving an impressive accuracy of 99.26% in cross‐validation across k ‐values of three, four and five folds. This research advances the field by employing multiclass classification for dysgraphia detection and severity assessment, offering a pathway to personalized interventions tailored to the needs of affected children.
Ramlan et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: