Pathologies affect the performance of ECG signals compression

Area of Science:

• Biomedical Engineering
• Signal Processing
• Medical Informatics

Background:

• Electrocardiogram (ECG) signal compression is crucial for data storage and transmission.
• The influence of ECG pathologies on compression performance remains understudied.
• Existing research lacks focus on how ECG abnormalities affect compression algorithms.

Purpose of the Study:

• To investigate the impact of ECG pathologies on the efficiency and quality of compression algorithms.
• To evaluate whether signal rhythm and morphology affect ECG compression performance.
• To provide novel annotations for ECG signal pathologies in a standard database.

Main Methods:

• Compression of 125 15-lead ECG signals from the CSE database using two algorithms: single-cycle fractal-based and wavelet transform with set partitioning in hierarchical trees.
• Annotation of ECG signals' rhythm and morphology as either physiological or pathological.
• Statistical evaluation of compression performance metrics.

Main Results:

• Physiological ECG signals were compressed with significantly better quality than pathological signals across multiple metrics.
• Pathological signals exhibited lower compression efficiency compared to physiological signals.
• ECG signals with both pathological rhythm and morphology showed the poorest compression outcomes.

Conclusions:

• ECG pathologies demonstrably affect the performance of compression algorithms, reducing both quality and efficiency.
• The presence of abnormal rhythm and morphology in ECG signals poses challenges for effective data compression.
• This study provides the first evidence of ECG pathology's impact on compression and releases new annotated data.