Untrimmed Video Classification for Activity Detection: submission to ActivityNet Challenge
arXiv:1607.01979
Abstract
Current state-of-the-art human activity recognition is focused on the classification of temporally trimmed videos in which only one action occurs per frame. We propose a simple, yet effective, method for the temporal detection of activities in temporally untrimmed videos with the help of untrimmed classification. Firstly, our model predicts the top k labels for each untrimmed video by analysing global video-level features. Secondly, frame-level binary classification is combined with dynamic programming to generate the temporally trimmed activity proposals. Finally, each proposal is assigned a label based on the global label, and scored with the score of the temporal activity proposal and the global score. Ultimately, we show that untrimmed video classification models can be used as stepping stone for temporal detection.
3 pages, Presented at ActivityNet Large Scale Activity Recognition Challenge workshop at CVPR 2016, Second position in ActivityNet Detection challenge 2016
Cited by in corpus (14)
- A Pursuit of Temporal Accuracy in General Activity Detection
- BMN: Boundary-Matching Network for Temporal Action Proposal Generation
- Temporal Context Network for Activity Localization in Videos
- Temporal Action Proposal Generation with Transformers
- Temporal Context Aggregation Network for Temporal Action Proposal Refinement
- AdapNet: Adaptability Decomposing Encoder-Decoder Network for Weakly Supervised Action Recognition and Localization
- Learning to Localize Actions from Moments
- Two-Stream Region Convolutional 3D Network for Temporal Activity Detection
- Multi-Level Temporal Pyramid Network for Action Detection
- Transferable Knowledge-Based Multi-Granularity Aggregation Network for Temporal Action Localization: Submission to ActivityNet Challenge 2021
- Learning Temporal Action Proposals With Fewer Labels
- Temporal Action Localization using Long Short-Term Dependency
- Online Spatiotemporal Action Detection and Prediction via Causal Representations
- Complementary Boundary Generator with Scale-Invariant Relation Modeling for Temporal Action Localization: Submission to ActivityNet Challenge 2020