3D Convolutional Neural Networks for Human Action Recognition

TLDR

Human action recognition in surveillance videos traditionally relies on handcrafted features, while CNNs can process raw inputs but are limited to 2D, motivating the need for 3D approaches. The paper proposes a novel 3D CNN for action recognition. The 3D CNN extracts spatial and temporal features via 3D convolutions, generating multiple channels whose representations are fused, and its performance is further enhanced by regularizing outputs with high‑level features and ensembling diverse models. Applied to airport surveillance videos, the models outperform baseline methods.

Abstract

We consider the automated recognition of human actions in surveillance videos. Most current methods build classifiers based on complex handcrafted features computed from the raw inputs. Convolutional neural networks (CNNs) are a type of deep model that can act directly on the raw inputs. However, such models are currently limited to handling 2D inputs. In this paper, we develop a novel 3D CNN model for action recognition. This model extracts features from both the spatial and the temporal dimensions by performing 3D convolutions, thereby capturing the motion information encoded in multiple adjacent frames. The developed model generates multiple channels of information from the input frames, and the final feature representation combines information from all channels. To further boost the performance, we propose regularizing the outputs with high-level features and combining the predictions of a variety of different models. We apply the developed models to recognize human actions in the real-world environment of airport surveillance videos, and they achieve superior performance in comparison to baseline methods.