Speeding up an Adaptive Filter based ECG Signal Pre-processing on Embedded Architectures

Medical applications increasingly require complex calculations with constraints of accelerated processing time. These are therefore oriented towards the integration high-performance embedded architectures. In this context, detection cardiac abnormalities is a task that remains high priority in emergency medicine. ECG analysis requires significant computing time since large amount information must be analyzed parallel frequencies. Real-time biggest challenge for researchers, when talking about like activity monitoring. This work evaluates Adaptive Dual Threshold Filter (ADTF) algorithm dedicated to signal filtering using various architectures: A Raspberry 3B+ and Odroid XU4. The implementation has been based on C/C++ OpenMP exploit parallelism used evaluation was validated several signals proposed MIT-BIH Arrhythmia database sampling frequency 360 Hz. Based an algorithmic complexity study parallelization functional blocks which present workloads, results show mean execution 7.5 ms 0.34 With efficient XU4 architecture, real-time performance can achieved.