Time Frequency noise canceller for an optimized separation of the ECG from low back sEMG signals

This work describes a hybrid adaptive filtering technique, designed to optimize the removal of the electrocardiogram (ECG) from surface electromyography (sEMG) recordings. Using the adaptive noise cancellation technique (ANC), we minimize the error gradient based on combined Time-Frequency filtering enhancements. The validation on artificially prepared signals with a real data basis shows better frequency and time filtering performances in comparison to simulations with standard implementations.