Relation Modeling in Spatio-Temporal Action Localization
arXiv:2106.08061
Abstract
This paper presents our solution to the AVA-Kinetics Crossover Challenge of ActivityNet workshop at CVPR 2021. Our solution utilizes multiple types of relation modeling methods for spatio-temporal action detection and adopts a training strategy to integrate multiple relation modeling in end-to-end training over the two large-scale video datasets. Learning with memory bank and finetuning for long-tailed distribution are also investigated to further improve the performance. In this paper, we detail the implementations of our solution and provide experiments results and corresponding discussions. We finally achieve 40.67 mAP on the test set of AVA-Kinetics.
CVPR 2021 ActivityNet Workshop Report
References in corpus (8)
- Semi-Supervised Classification with Graph Convolutional Networks
- Decoupling Representation and Classifier for Long-Tailed Recognition
- The AVA-Kinetics Localized Human Actions Video Dataset
- Multiscale Vision Transformers
- Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection
- Temporal Context Aggregation Network for Temporal Action Proposal Refinement
- Self-supervised Motion Learning from Static Images
- CBR-Net: Cascade Boundary Refinement Network for Action Detection: Submission to ActivityNet Challenge 2020 (Task 1)