BMN: Boundary-Matching Network for Temporal Action Proposal Generation

TL;DR

This paper introduces BMN, a boundary-matching network that efficiently generates precise temporal action proposals with reliable confidence scores, significantly improving performance on challenging datasets.

Contribution

The paper proposes the Boundary-Matching mechanism and an end-to-end network, BMN, for improved temporal action proposal generation with accurate boundaries and confidence scores.

Findings

Significant performance improvements on THUMOS-14 and ActivityNet-1.3 datasets.

Efficient end-to-end training of proposal generation and confidence scoring.

Achieves state-of-the-art temporal action detection when combined with classifiers.

Abstract

Temporal action proposal generation is an challenging and promising task which aims to locate temporal regions in real-world videos where action or event may occur. Current bottom-up proposal generation methods can generate proposals with precise boundary, but cannot efficiently generate adequately reliable confidence scores for retrieving proposals. To address these difficulties, we introduce the Boundary-Matching (BM) mechanism to evaluate confidence scores of densely distributed proposals, which denote a proposal as a matching pair of starting and ending boundaries and combine all densely distributed BM pairs into the BM confidence map. Based on BM mechanism, we propose an effective, efficient and end-to-end proposal generation method, named Boundary-Matching Network (BMN), which generates proposals with precise temporal boundaries as well as reliable confidence scores simultaneously. The two-branches of BMN are jointly trained in an unified framework. We conduct experiments on two challenging datasets: THUMOS-14 and ActivityNet-1.3, where BMN shows significant performance improvement with remarkable efficiency and generalizability. Further, combining with existing action classifier, BMN can achieve state-of-the-art temporal action detection performance.