An Image is Worth 16x16 Words, What is a Video Worth?

Abstract

Leading methods in the domain of action recognition try to distill information from both the spatial and temporal dimensions of an input video. Methods that reach State of the Art (SotA) accuracy, usually make use of 3D convolution layers as a way to abstract the temporal information from video frames. The use of such convolutions requires sampling short clips from the input video, where each clip is a collection of closely sampled frames. Since each short clip covers a small fraction of an input video, multiple clips are sampled at inference in order to cover the whole temporal length of the video. This leads to increased computational load and is impractical for real-world applications. We address the computational bottleneck by significantly reducing the number of frames required for inference. Our approach relies on a temporal transformer that applies global attention over video frames, and thus better exploits the salient information in each frame. Therefore our approach is very input efficient, and can achieve SotA results (on Kinetics dataset) with a fraction of the data (frames per video), computation and latency. Specifically on Kinetics-400, we reach $80.5$ top-1 accuracy with $\times 30$ less frames per video, and $\times 40$ faster inference than the current leading method. Code is available at: https://github.com/Alibaba-MIIL/STAM

References in corpus (6)

• Two-Stream Convolutional Networks for Action Recognition in Videos
• UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
• The Kinetics Human Action Video Dataset
• Stand-Alone Self-Attention in Vision Models
• Dynamic Sampling Networks for Efficient Action Recognition in Videos
• Quantifying Translation-Invariance in Convolutional Neural Networks

Cited by in corpus (10)

• VATT: Transformers for Multimodal Self-Supervised Learning from Raw Video, Audio and Text
• ActionCLIP: A New Paradigm for Video Action Recognition
• CLIP2Video: Mastering Video-Text Retrieval via Image CLIP
• ATISS: Autoregressive Transformers for Indoor Scene Synthesis
• Long Short-Term Transformer for Online Action Detection
• FastLTS: Non-Autoregressive End-to-End Unconstrained Lip-to-Speech Synthesis
• ImageNet-21K Pretraining for the Masses
• Astronomical image time series classification using CONVolutional attENTION (ConvEntion)
• Evaluating Transformers for Lightweight Action Recognition
• Long-Short Temporal Contrastive Learning of Video Transformers