Liao, S; Lyons, T; Yang, W; Schlegel, K; Ni, H; (2021) Logsig-RNN: a novel network for robust and efficient skeleton-based action recognition. In: Proceedings of the The British Machine Vision Conference (BMVC) 2021. The British Machine Vision Association: Virtual Conference.

Preview

Text 2110.13008v2.pdf - Accepted Version Download (1MB) | Preview

Abstract

This paper contributes to the challenge of skeleton-based human action recognition in videos. The key step is to develop a generic network architecture to extract discriminative features for the spatio-temporal skeleton data. In this paper, we propose a novel module, namely Logsig-RNN, which is the combination of the log-signature layer and recurrent type neural networks (RNNs). The former one comes from the mathematically principled technology of signatures and log-signatures as representations for streamed data, which can manage high sample rate streams, non-uniform sampling and time series of variable length. It serves as an enhancement of the recurrent layer, which can be conveniently plugged into neural networks. Besides we propose two path transformation layers to significantly reduce path dimension while retaining the essential information fed into the Logsig-RNN module. (The network architecture is illustrated in Figure 1 (Right).) Finally, numerical results demonstrate that replacing the RNN module by the LogsigRNN module in SOTA networks consistently improves the performance on both Chalearn gesture data and NTU RGB+D 120 action data in terms of accuracy and robustness. In particular, we achieve the state-of-the-art accuracy on Chalearn2013 gesture data by combining simple path transformation layers with the Logsig-RNN.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item