Neurodegenerative disorders [12], which include a wide range of complex and currently incurable conditions, pose major challenges to modern healthcare. This study emphasizes the early identification [18] of Parkinson’s disease [14] through offline handwriting analysis, a relatively novel method in clinical diagnostics. To tackle this issue, we introduce Cure-Net, a convolutional neural network framework aimed at enhancing both the reliability and precision of Parkinson’s disease detection. Detecting the disease at an early stage is vital for improving patient management, treatment strategies, and overall quality of life. Extensive evaluation of Cure-Net on handwriting-based datasets demonstrated its effectiveness. The model achieved a balanced accuracy of 92.57% on the Spiral–Meander combination in the HandPD dataset, while also reaching 98.05% accuracy on the Meander task and 97.34% on another Meander task from the NewHandPD dataset. These outcomes showcase Cure-Net’s strong potential for accurate and early diagnosis, outperforming previously reported methods. Such advancements mark a meaningful step toward more effective diagnostic solutions and better patient outcomes.
