Quality Enhancement of Compressed Vibrotactile Signals Using Recurrent Neural Networks and Residual Learning

• Published in: IEEE transactions on haptics Vol. 14; no. 2; pp. 316 - 321
• Main Authors: Noll, Andreas, Gurbuz, Ayten, Gulecyuz, Basak, Cui, Kai, Steinbach, Eckehard
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.04.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Ablation; Artificial neural networks; Codecs; Compression ratio; Learning; machine learning; Neural networks; Quality enhancement; Recurrent neural networks; residual learning; RNN; Signal processing; Signal quality; Signal to noise ratio; tactile signal compression; Task analysis; Testing; Training
• ISSN: 1939-1412; 2329-4051
• DOI: 10.1109/TOH.2021.3078889

We present a neural network-based compression artifact removal technique for vibrotactile signals. The proposed decoder-side quality enhancement approach is based on recurrent neural networks (RNNs) and the principle of residual learning. We use a total of 8 nonlinear RNN layers trained to first estimate the difference between the original and the compressed signal. The estimated difference signal is then added to the compressed signal, followed by further linear processing steps to construct the enhanced signal. With our approach, we are able to enhance signals at almost all compression ratios by up to 1.25\ \mathrm {dB}. For the signals in our data set, rougly 86% are enhanced in their quality. Through an ablation study, we show that every block of our network is functioning as intended and contributes to the compression artifact removal. Additionally, we show that the chosen network parameters maximize performance.