A Real Time Wavelet Filtering for ECG Baseline Wandering Removal

Electrocardiogram (ECG) signal analysis plays an unparalleled role in diagnosis of heart diseases. The raw ECG signal usually consists of low-frequency baseline wander (BW) which must be removed for a better diagnosis. This paper presents a real-time implementation of the Discrete Wavelet Transform based filtering approach for BW removal from ECG signal (obtained from MIT-BIH database) using a 16-bit, fixed-point DSP board (TMS320C5515). The results indicate that the wavelet-based filter was effective in removing BW and was efficient for a real-time implementation, when compared with the traditional FIR and IIR filters. These results were compared and validated with the offline implementations using MATLAB.