Attention Clusters: Purely Attention Based Local Feature Integration for Video Classification

TL;DR

This paper introduces a purely attention-based local feature integration framework for video classification, demonstrating competitive results without relying on traditional CNNs or RNNs, and achieving state-of-the-art performance on large datasets.

Contribution

The paper proposes a novel attention clusters framework with a shifting operation for local feature integration in video classification, challenging the necessity of long-term temporal modeling.

Findings

Achieves 79.4% top-1 accuracy on Kinetics dataset

Outperforms many existing methods in video classification

Wins the ActivityNet Kinetics Challenge 2017

Abstract

Recently, substantial research effort has focused on how to apply CNNs or RNNs to better extract temporal patterns from videos, so as to improve the accuracy of video classification. In this paper, however, we show that temporal information, especially longer-term patterns, may not be necessary to achieve competitive results on common video classification datasets. We investigate the potential of a purely attention based local feature integration. Accounting for the characteristics of such features in video classification, we propose a local feature integration framework based on attention clusters, and introduce a shifting operation to capture more diverse signals. We carefully analyze and compare the effect of different attention mechanisms, cluster sizes, and the use of the shifting operation, and also investigate the combination of attention clusters for multimodal integration. We demonstrate the effectiveness of our framework on three real-world video classification datasets. Our model achieves competitive results across all of these. In particular, on the large-scale Kinetics dataset, our framework obtains an excellent single model accuracy of 79.4% in terms of the top-1 and 94.0% in terms of the top-5 accuracy on the validation set. The attention clusters are the backbone of our winner solution at ActivityNet Kinetics Challenge 2017. Code and models will be released soon.