Scholarlink

Diagnosis of Parkinson’s disease early and accurately continues to be a significant challenge. This study investigates the utilization of deep learning models for the diagnosis of Parkinson’s disease through MRI-based analysis. A novel 3D implementation of the residual Kolmogorov-Arnold Network is introduced, which integrates RKAN blocks into variations of DenseNet and ResNet architectures on the PPMI dataset. These hybrid models are designed to enhance feature extraction and classification capabilities, effectively addressing the complexities associated with 3D medical imaging data. RKANs, when integrated with DenseNet-201, demonstrated superior performance across diverse datasets and dimensionalities, achieving the highest accuracy, with ResNet-50 showing comparable performance.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Parkinson’s Disease Diagnosis via KANs: A Deep Learning Perspective

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Krupal Lad,
  • Shivani Desai,
  • Tarjni Vyas,
  • Anuja Nair,
  • Lata Gohil,
  • Sheshang Degadwala

摘要

Diagnosis of Parkinson’s disease early and accurately continues to be a significant challenge. This study investigates the utilization of deep learning models for the diagnosis of Parkinson’s disease through MRI-based analysis. A novel 3D implementation of the residual Kolmogorov-Arnold Network is introduced, which integrates RKAN blocks into variations of DenseNet and ResNet architectures on the PPMI dataset. These hybrid models are designed to enhance feature extraction and classification capabilities, effectively addressing the complexities associated with 3D medical imaging data. RKANs, when integrated with DenseNet-201, demonstrated superior performance across diverse datasets and dimensionalities, achieving the highest accuracy, with ResNet-50 showing comparable performance.