A Pursuit of Temporal Accuracy in General Activity Detection

Abstract

Detecting activities in untrimmed videos is an important but challenging task. The performance of existing methods remains unsatisfactory, e.g., they often meet difficulties in locating the beginning and end of a long complex action. In this paper, we propose a generic framework that can accurately detect a wide variety of activities from untrimmed videos. Our first contribution is a novel proposal scheme that can efficiently generate candidates with accurate temporal boundaries. The other contribution is a cascaded classification pipeline that explicitly distinguishes between relevance and completeness of a candidate instance. On two challenging temporal activity detection datasets, THUMOS14 and ActivityNet, the proposed framework significantly outperforms the existing state-of-the-art methods, demonstrating superior accuracy and strong adaptivity in handling activities with various temporal structures.

References in corpus (4)

• UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Untrimmed Video Classification for Activity Detection: submission to ActivityNet Challenge
• Temporal Activity Detection in Untrimmed Videos with Recurrent Neural Networks

Cited by in corpus (14)

• Revisiting Anchor Mechanisms for Temporal Action Localization
• BMN: Boundary-Matching Network for Temporal Action Proposal Generation
• Look Closer to Ground Better: Weakly-Supervised Temporal Grounding of Sentence in Video
• Weakly Supervised Temporal Adjacent Network for Language Grounding
• Self-Supervised Learning for Semi-Supervised Temporal Action Proposal
• Boundary Content Graph Neural Network for Temporal Action Proposal Generation
• A Self-Adaptive Proposal Model for Temporal Action Detection based on Reinforcement Learning
• Learning to Localize Actions from Moments
• Multi-Level Temporal Pyramid Network for Action Detection
• Class Semantics-based Attention for Action Detection
• CMSN: Continuous Multi-stage Network and Variable Margin Cosine Loss for Temporal Action Proposal Generation
• Temporal Action Localization using Long Short-Term Dependency
• Complementary Boundary Generator with Scale-Invariant Relation Modeling for Temporal Action Localization: Submission to ActivityNet Challenge 2020
• TruNet: Short Videos Generation from Long Videos via Story-Preserving Truncation