An FPGA Based Tracking Implementation for Parkinson's Patients.

Area of Science:

• Biomedical Engineering
• Computer Vision
• Medical Technology

Background:

• Parkinson's disease patient tracking requires efficient and low-power computer vision systems.
• Existing systems often have high computational demands and hardware requirements.
• Optimization is needed for real-world clinical applications.

Purpose of the Study:

• To optimize a computer vision-based tracking system for Parkinson's disease patients.
• To improve system efficiency in terms of energy consumption and hardware needs.
• To implement the system on a low-cost development platform for practical use.

Main Methods:

• Optimized background subtraction using Gaussian Mixture Models (GMM) for frame segmentation.
• Implemented the GMM module on a ZedBoard featuring Xilinx Zynq XC7Z020 SoC (System on Chip).
• Combined ARM Processor and FPGA for efficient computation.

Main Results:

• Achieved significant improvements in system efficiency and reduced hardware requirements.
• Demonstrated real-time performance with low power consumption.
• Validated accurate tracking in real medical settings.

Conclusions:

• The optimized system offers an efficient, low-cost solution for Parkinson's disease patient tracking.
• The use of GMM on an SoC platform enables accurate and real-time monitoring.
• This approach is suitable for deployment in day hospital centers and similar clinical environments.