Scholarlink

This project helps to diagnose Parkinson’s disease (PD) in the early stage by using different computer models. All of these models are easy to understand and trustworthy. Based on the spiral handwritten photographs, this study offers a thorough assessment of many pre-trained convolutional neural network (CNN) models for categorization of PD. We have used DenseNet-201, VGG-19, RedNet-50, and InceptionV3 architectures. Also, key performance measures like F1-score, recall, accuracy, and precision are contrasted. DenseNet-201 came out as optimal model with the accuracy of 95.45% and minimal error rate 4.55%. DenseNet emerged as most efficient model outperforming all other CNN models. This study is about medical picture analysis and predicting the disease by providing the details to the deep learning models to diagnose.

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

Detecting Parkinson’s Disease Early: A Study of Spiral Handwriting Utilizing Pre-trained CNN Models

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Rushali Deshmukh,
  • Shreyash Kudale,
  • Satyam Pawar,
  • Shweta Jagadale,
  • Harshal Tambe

摘要

This project helps to diagnose Parkinson’s disease (PD) in the early stage by using different computer models. All of these models are easy to understand and trustworthy. Based on the spiral handwritten photographs, this study offers a thorough assessment of many pre-trained convolutional neural network (CNN) models for categorization of PD. We have used DenseNet-201, VGG-19, RedNet-50, and InceptionV3 architectures. Also, key performance measures like F1-score, recall, accuracy, and precision are contrasted. DenseNet-201 came out as optimal model with the accuracy of 95.45% and minimal error rate 4.55%. DenseNet emerged as most efficient model outperforming all other CNN models. This study is about medical picture analysis and predicting the disease by providing the details to the deep learning models to diagnose.