Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Exploring the Efficacy of Modified Transfer Learning in Identifying Parkinson's Disease Through Drawn Image Patterns

Nabil Daiyan, Md Rakibul Haque

TL;DR

This work addresses the challenge of early, non-invasive Parkinson's disease diagnosis by leveraging hand-drawn spiral and wave drawings as biomarkers. It introduces a three-phase pipeline that combines pre-trained CNN feature extractors (VGG16/VGG19), added convolutional and attention layers, and hard-voting ensembles to detect PD patterns. The spiral and wave classifiers achieve 90% and 96.67% accuracy respectively, with a final ensemble accuracy of 93.3% on unaugmented data and 98% after fusion, indicating robust performance and potential clinical utility. The approach offers a scalable, non-invasive screening tool that could complement clinical assessments and facilitate earlier intervention, pending validation on larger datasets.

Abstract

Parkinson's disease (PD) is a progressive neurodegenerative condition characterized by the death of dopaminergic neurons, leading to various movement disorder symptoms. Early diagnosis of PD is crucial to prevent adverse effects, yet traditional diagnostic methods are often cumbersome and costly. In this study, a machine learning-based approach is proposed using hand-drawn spiral and wave images as potential biomarkers for PD detection. Our methodology leverages convolutional neural networks (CNNs), transfer learning, and attention mechanisms to improve model performance and resilience against overfitting. To enhance the diversity and richness of both spiral and wave categories, the training dataset undergoes augmentation to increase the number of images. The proposed architecture comprises three phases: utilizing pre-trained CNNs, incorporating custom convolutional layers, and ensemble voting. Employing hard voting further enhances performance by aggregating predictions from multiple models. Experimental results show promising accuracy rates. For spiral images, weighted average precision, recall, and F1-score are 90%, and for wave images, they are 96.67%. After combining the predictions through ensemble hard voting, the overall accuracy is 93.3%. These findings underscore the potential of machine learning in early PD diagnosis, offering a non-invasive and cost-effective solution to improve patient outcomes.

Exploring the Efficacy of Modified Transfer Learning in Identifying Parkinson's Disease Through Drawn Image Patterns

TL;DR

This work addresses the challenge of early, non-invasive Parkinson's disease diagnosis by leveraging hand-drawn spiral and wave drawings as biomarkers. It introduces a three-phase pipeline that combines pre-trained CNN feature extractors (VGG16/VGG19), added convolutional and attention layers, and hard-voting ensembles to detect PD patterns. The spiral and wave classifiers achieve 90% and 96.67% accuracy respectively, with a final ensemble accuracy of 93.3% on unaugmented data and 98% after fusion, indicating robust performance and potential clinical utility. The approach offers a scalable, non-invasive screening tool that could complement clinical assessments and facilitate earlier intervention, pending validation on larger datasets.

Abstract

Parkinson's disease (PD) is a progressive neurodegenerative condition characterized by the death of dopaminergic neurons, leading to various movement disorder symptoms. Early diagnosis of PD is crucial to prevent adverse effects, yet traditional diagnostic methods are often cumbersome and costly. In this study, a machine learning-based approach is proposed using hand-drawn spiral and wave images as potential biomarkers for PD detection. Our methodology leverages convolutional neural networks (CNNs), transfer learning, and attention mechanisms to improve model performance and resilience against overfitting. To enhance the diversity and richness of both spiral and wave categories, the training dataset undergoes augmentation to increase the number of images. The proposed architecture comprises three phases: utilizing pre-trained CNNs, incorporating custom convolutional layers, and ensemble voting. Employing hard voting further enhances performance by aggregating predictions from multiple models. Experimental results show promising accuracy rates. For spiral images, weighted average precision, recall, and F1-score are 90%, and for wave images, they are 96.67%. After combining the predictions through ensemble hard voting, the overall accuracy is 93.3%. These findings underscore the potential of machine learning in early PD diagnosis, offering a non-invasive and cost-effective solution to improve patient outcomes.

Paper Structure

This paper contains 16 sections, 2 equations, 7 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Methodology
  4. Dataset Description
  5. Transfer Learning
  6. Convolutional Neural Network
  7. Attention Mechanism
  8. Proposed Approach
  9. Data Preprocessing
  10. Data Augmentation
  11. Model Development
  12. Experimental Analysis
  13. Experimental Setup
  14. Result Analysis
  15. Performance Analysis
  16. ...and 1 more sections

Figures (7)

  • Figure 1: Spiral and Wave Drawings for Healthy and Parkinson's Disease
  • Figure 2: Sample of spiral and wave drawing after image preprocessing
  • Figure 3: Data distribution after image augmentation
  • Figure 4: Proposed model architecture
  • Figure 5: Accuracy and loss curve for spiral drawings
  • ...and 2 more figures