Multi-Level Recurrent Residual Networks for Action Recognition

TL;DR

This paper introduces Multi-Level Recurrent Residual Networks (MRRN), a novel architecture combining residual networks and recurrent units across multiple levels to improve action recognition by effectively modeling spatiotemporal features with lower complexity.

Contribution

The paper proposes a multi-stream, multi-level recurrent residual network that captures diverse spatiotemporal features efficiently and outperforms previous CNN-RNN models in action recognition tasks.

Findings

Achieves 51.3% accuracy on HMDB-51 dataset.

Achieves 81.9% accuracy on UCF-101 dataset.

Lower complexity compared to previous models.

Abstract

Most existing Convolutional Neural Networks(CNNs) used for action recognition are either difficult to optimize or underuse crucial temporal information. Inspired by the fact that the recurrent model consistently makes breakthroughs in the task related to sequence, we propose a novel Multi-Level Recurrent Residual Networks(MRRN) which incorporates three recognition streams. Each stream consists of a Residual Networks(ResNets) and a recurrent model. The proposed model captures spatiotemporal information by employing both alternative ResNets to learn spatial representations from static frames and stacked Simple Recurrent Units(SRUs) to model temporal dynamics. Three distinct-level streams learned low-, mid-, high-level representations independently are fused by computing a weighted average of their softmax scores to obtain the complementary representations of the video. Unlike previous models which boost performance at the cost of time complexity and space complexity, our models have a lower complexity by employing shortcut connection and are trained end-to-end with greater efficiency. MRRN displays significant performance improvements compared to CNN-RNN framework baselines and obtains comparable performance with the state-of-the-art, achieving 51.3% on HMDB-51 dataset and 81.9% on UCF-101 dataset although no additional data.