1st place solution for AVA-Kinetics Crossover in AcitivityNet Challenge 2020
Published in
arXiv.org(2020)External Links:
TL
TL;DR
This technical report introduces the winning solution to the spatio-temporal action localization track, AVA-Kinetics Crossover, in ActivityNet Challenge 2020, based on Actor-Context-Actor Relation Network, which outperforms other entries by a large margin.
Abstract
This technical report introduces our winning solution to the spatio-temporal action localization track, AVA-Kinetics Crossover, in ActivityNet Challenge 2020. Our entry is mainly based on Actor-Context-Actor Relation Network. We describe technical details for the new AVA-Kinetics dataset, together with some experimental results. Without any bells and whistles, we achieved 39.62 mAP on the test set of AVA-Kinetics, which outperforms other entries by a large margin. Code will be available at: this https URL.
References19 items
1
Actor-Context-Actor Relation Network for Spatio-Temporal Action Localization
2
The AVA-Kinetics Localized Human Actions Video Dataset
3
Revisiting the Sibling Head in Object Detector
4
1st Place Solutions for OpenImage2019 - Object Detection and Instance Segmentation
5