Task-Dependent Compression Performance Assessment in Body Physiology Signals

• Published in: 2020 28th Signal Processing and Communications Applications Conference (SIU) pp. 1 - 4
• Main Authors: Ileri, Fatih, Gokcay, Didem
• Format: Conference Proceeding
• Language: English
• Published: IEEE 05.10.2020
• Subjects: DCT; Discrete cosine transforms; Electroencephalography; Electromyography; entropy coding; Measurement; Multilayer perceptrons; Neural networks; signal compression; Task analysis
• DOI: 10.1109/SIU49456.2020.9302049

Compression of biological signals have great importance in storing, offline processing, and transmission of these signals to other entities. There have been many researches based on assessing the performance of newly published compression methods. Main performance metrics used in these researches are the compression ratio (CR) and the error metric, generally the PRD (percent mean square difference). In this research we assess the performance of our compression method in a task-dependent fashion. We applied a neural network based classification on original and compressed EMG signals, collected from subjects participated a defined task. With this approach, we were able to understand what amount of necessary information of EMG signals is suppressed during compression.