Abstract—In this study, we propose a model for determining depression severity through handwriting analysis. Traditional diagnostic methods rely on subjective communication between doctor and patient, which can lead to varied interpretations and prolonged diagnostic timelines. It is expected that handwritten data can be used to objectively diagnose depression. We utilize Residual Neural Network (ResNet) and Gradient-weighted Class Activation Mapping++ (Grad-CAM++) to identify regions of interest in handwriting images, associating these with normalized handwriting speed, which has been shown to correlate with depressive states. Experimental results showed that the model’s region of interest focuses on the slow rather than the fast rate part. This model approach facilitates early and efficient detection of depression by making the process more accessible and minimizing the need for specialized equipment.
 
Keywords—depression, handwriting analysis, Convolutional Neural Network (CNN), Gradient-weighted Class Activation Mapping++ (Grad-CAM++), Residual Neural Network (ResNet)


Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).