Contrast and Homogeneity Feature Analysis for Classifying Tremor Levels in Parkinson's Disease Patients

Area of Science:

• Neurology
• Biomedical Engineering
• Machine Learning

Background:

• Early detection of Parkinson's disease (PD) is crucial for effective treatment and improved patient quality of life.
• Tremors are a key symptom of PD, and assessing their severity aids in diagnosis and management.
• Current diagnostic methods may not always capture subtle tremor variations indicative of early-stage PD.

Purpose of the Study:

• To propose a non-invasive strategy for measuring tremor severity to diagnose early-stage Parkinson's disease.
• To classify tremors into different levels based on the Unified Parkinson's Disease Rating Scale (UPDRS).
• To evaluate the effectiveness of a machine learning approach using a Leap Motion Controller for tremor analysis.

Main Methods:

• Utilized a Leap Motion Controller to capture 3D coordinates of hand and finger movements.
• Computed texture features using Sum and Difference Histograms (SDH) from the acquired data.
• Employed a machine learning classifier for the final classification of tremor severity levels.

Main Results:

• The proposed non-invasive method demonstrated effectiveness in measuring tremor severity.
• Texture features derived from hand movements proved valuable for tremor characterization.
• The machine learning classifier achieved accurate classification of different tremor levels.

Conclusions:

• The developed non-invasive strategy offers a promising tool for the early diagnosis of Parkinson's disease.
• This approach can potentially enhance the accuracy of PD staging and inform therapeutic interventions.
• Further research can explore refining the feature extraction and classification models for broader clinical application.