|
|
Compression and Quality Assessment of ECG Signals
Němcová, Andrea ; Tkacz,, Professor Ewaryst (oponent) ; Kudrna,, Petr (oponent) ; Vítek, Martin (vedoucí práce)
Lossy compression of ECG signals is useful and still challenging and developing area. In recent years, new and new compression algorithms are developed. Unfortunately, there is a lack of standards for compression quality assessment. Thus, there are plenty of methods, performance of which cannot be objectively compared at all or can be compared only in a rough guess. Moreover, it is not much known, whether and how the pathologies in ECG signals influence the compression algorithms performance. In this thesis, a review of all found methods for the assessment of ECG signal quality after compression and reconstruction is presented. 10 new methods were created. The known and the new methods were analysed and based on the results, 12 of them were recommended for further use. New Single-Cycle Fractal-Based (SCyF) compression algorithm is introduced. SCyF algorithm was inspired by fractal-based method and uses one cycle of ECG signal as a domain. It was tested on four different databases using 12 recommended quality metrics and compared with known and popular method based on wavelet transform and Set Partitioning in Hierarchical Trees (SPIHT). The testing process serves as an example how the standardization of ECG signal compression assessment should look like. Further, it was statistically proven that the difference between compression of physiological and pathological signals exists. Pathological signals were compressed with lower efficiency and quality than physiological signals.
|
|
|
Compression and Quality Assessment of ECG Signals
Němcová, Andrea ; Tkacz,, Professor Ewaryst (oponent) ; Kudrna,, Petr (oponent) ; Vítek, Martin (vedoucí práce)
Lossy compression of ECG signals is useful and still challenging and developing area. In recent years, new and new compression algorithms are developed. Unfortunately, there is a lack of standards for compression quality assessment. Thus, there are plenty of methods, performance of which cannot be objectively compared at all or can be compared only in a rough guess. Moreover, it is not much known, whether and how the pathologies in ECG signals influence the compression algorithms performance. In this thesis, a review of all found methods for the assessment of ECG signal quality after compression and reconstruction is presented. 10 new methods were created. The known and the new methods were analysed and based on the results, 12 of them were recommended for further use. New Single-Cycle Fractal-Based (SCyF) compression algorithm is introduced. SCyF algorithm was inspired by fractal-based method and uses one cycle of ECG signal as a domain. It was tested on four different databases using 12 recommended quality metrics and compared with known and popular method based on wavelet transform and Set Partitioning in Hierarchical Trees (SPIHT). The testing process serves as an example how the standardization of ECG signal compression assessment should look like. Further, it was statistically proven that the difference between compression of physiological and pathological signals exists. Pathological signals were compressed with lower efficiency and quality than physiological signals.
|
host ::
RSS.