Sparse Analyzer Tool for Biomedical Signals.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Data Science

Background:

• Incomplete data acquisition in biomedical sensing leads to signal damage, necessitating robust recovery methods.
• Compressive Sensing (CS) offers a framework for reconstructing signals from under-sampled data, crucial for various medical applications.
• Existing CS reconstruction algorithms vary in effectiveness depending on signal type and data completeness.

Purpose of the Study:

• To present a virtual instrument for evaluating and comparing CS reconstruction algorithms for biomedical signals.
• To provide a statistical analysis tool for assessing algorithm performance under varying degrees of data missingness.
• To facilitate the selection of optimal CS algorithms for specific biomedical applications and signal types.

Main Methods:

• Implementation of various CS reconstruction algorithms within a software-based virtual instrument.
• Inclusion of a statistical analyzer to quantify reconstruction accuracy (e.g., Mean Square Error) and computational time.
• Capability to test algorithms with diverse biomedical signals and adjustable percentages of missing data, using internal or external datasets.

Main Results:

• The virtual instrument allows for quantitative comparison of different CS reconstruction algorithms.
• Performance metrics like Mean Square Error and computational time are used to evaluate algorithm efficiency.
• The study demonstrates the instrument's utility in identifying suitable CS approaches for under-sampled biomedical data.

Conclusions:

• The developed virtual instrument serves as a valuable tool for researchers and practitioners in biomedical signal processing.
• It enables informed selection of CS algorithms by providing objective performance analysis.
• This facilitates improved recovery of biomedical signals acquired with incomplete data, enhancing diagnostic capabilities.