Weakly Supervised Action Localization by Sparse Temporal Pooling Network

TL;DR

This paper introduces a weakly supervised method for localizing actions in untrimmed videos by identifying key segments with an attention mechanism and adaptive pooling, achieving state-of-the-art results without needing detailed annotations.

Contribution

It presents a novel sparse temporal pooling network that learns to identify key action segments using only video-level labels, improving localization accuracy.

Findings

Achieves state-of-the-art results on THUMOS14 dataset.

Outperforms previous weakly supervised methods on ActivityNet1.3.

Effectively localizes actions with minimal supervision.

Abstract

We propose a weakly supervised temporal action localization algorithm on untrimmed videos using convolutional neural networks. Our algorithm learns from video-level class labels and predicts temporal intervals of human actions with no requirement of temporal localization annotations. We design our network to identify a sparse subset of key segments associated with target actions in a video using an attention module and fuse the key segments through adaptive temporal pooling. Our loss function is comprised of two terms that minimize the video-level action classification error and enforce the sparsity of the segment selection. At inference time, we extract and score temporal proposals using temporal class activations and class-agnostic attentions to estimate the time intervals that correspond to target actions. The proposed algorithm attains state-of-the-art results on the THUMOS14 dataset and outstanding performance on ActivityNet1.3 even with its weak supervision.