Learning Spatiotemporal Features with 3D Convolutional Networks

TL;DR

This paper introduces 3D ConvNets for spatiotemporal video feature learning, demonstrating their superiority over 2D ConvNets, with simple architectures that outperform state-of-the-art methods on multiple benchmarks.

Contribution

It presents a straightforward 3D ConvNet architecture that effectively learns spatiotemporal features, outperforming existing methods and being computationally efficient and easy to train.

Findings

3D ConvNets outperform 2D ConvNets for spatiotemporal learning

Small 3x3x3 kernels in all layers yield top performance

C3D features achieve high accuracy with low-dimensional representations

Abstract

We propose a simple, yet effective approach for spatiotemporal feature learning using deep 3-dimensional convolutional networks (3D ConvNets) trained on a large scale supervised video dataset. Our findings are three-fold: 1) 3D ConvNets are more suitable for spatiotemporal feature learning compared to 2D ConvNets; 2) A homogeneous architecture with small 3x3x3 convolution kernels in all layers is among the best performing architectures for 3D ConvNets; and 3) Our learned features, namely C3D (Convolutional 3D), with a simple linear classifier outperform state-of-the-art methods on 4 different benchmarks and are comparable with current best methods on the other 2 benchmarks. In addition, the features are compact: achieving 52.8% accuracy on UCF101 dataset with only 10 dimensions and also very efficient to compute due to the fast inference of ConvNets. Finally, they are conceptually very simple and easy to train and use.