Neural network based canceller for Powerline Interference in ECG signals

Abstract

Power line interference may severely corrupt a biomedical recording. Notch Filters and adaptive chancellors have been suggested to suppress this interference. In this paper, an improved adaptive canceller for the reduction of the fundamental Power Line Interference component in electrocardiogram (EKG) recordings is proposed. A comparison is made between the performance of our method and a narrow and a wide Notch Filter and Notch Adaptive Filter in suppressing the fundamental power line interference component. For this purpose, a deal EGG signal is corrupted by an artificial power line interference signal. The cleaned signal after applying all methods is compared with the original EGG signal. Results indicate that power line interference of EKG are removed effectively by this new method. Interference elimination can be performed continuously and rapidly even if the situations of interference are changing with time or frequency. In the worst conditions 48.5Hz and 51.5Hz (BOW 1.5Hz), ANN obtained results show the efficiency (CCC=0.96 ± 0.02 SIR=17.3 ± 0.4) in comparison with the classical technique with the best performance (CCC=0.91±0.03 SIR=13.2±0.6). The method is easy to implement and it is applicable not only to EKG but also other biomedical signals.