Efficient thresholding-based ECG compressors for high quality applications using cosine modulated filter banks.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Data Compression

Background:

• Electrocardiogram (ECG) compression aims to reduce data size while preserving critical diagnostic information.
• Lossy thresholding compressors offer efficiency but may compromise signal fidelity.
• High-fidelity reconstruction is crucial for reliable medical applications.

Purpose of the Study:

• To propose and evaluate two novel ECG compression schemes.
• To assess their performance in scenarios demanding near-perfect signal reconstruction.
• To compare their effectiveness against existing discrete wavelet transform (DWT) methods.

Main Methods:

• Development of two compression algorithms utilizing nearly perfect reconstruction cosine modulated filter banks (CMFBs) for signal decomposition.
• Evaluation using the comprehensive MIT-BIH Arrhythmia Database.
• Application of appropriate metrics to quantify compression and reconstruction quality.

Main Results:

• The proposed CMFB-based compressors demonstrate superior performance compared to DWT-based techniques.
• This advantage is particularly evident under stringent high-quality reconstruction requirements.
• The methods achieve significant compression while maintaining signal integrity.

Conclusions:

• Cosine modulated filter banks provide an effective approach for high-fidelity ECG compression.
• The proposed schemes are well-suited for applications requiring minimal information loss.
• These CMFB compressors represent an advancement over current DWT-based compression strategies for critical ECG data.