Sparse Canonical Temporal Alignment With Deep Tensor Decomposition for Action Recognition

• Published in: IEEE transactions on image processing Vol. 26; no. 2; pp. 738 - 750
• Main Authors: Jia, Chengcheng, Shao, Ming, Fu, Yun
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.02.2017
• Subjects: Computers; deep non-negative tensor factorization; Feature extraction; Image recognition; key frames; Manifolds; Sparse canonical temporal alignment; Spatiotemporal phenomena; Tensile stress; tensor-train; Trajectory

In this paper, we solve three problems in action recognition: sub-action, multi-subject, and multi-modality, by reducing the diversity of intra-class samples. The main stage contains canonical temporal alignment and key frames selection. As we know, temporal alignment aims to reduce the diversity of...