English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Designing a real-time framework for the spatio-temporal action detection taskis still a challenge. In this paper, we propose a novel real-time actiondetection framework, YOWOv2. In this new framework, YOWOv2 takes advantage ofboth the 3D backbone and 2D backbone for accurate action detection. Amulti-level detection pipeline is designed to detect action instances ofdifferent scales. To achieve this goal, we carefully build a simple andefficient 2D backbone with a feature pyramid network to extract differentlevels of classification features and regression features. For the 3D backbone,we adopt the existing efficient 3D CNN to save development time. By combining3D backbones and 2D backbones of different sizes, we design a YOWOv2 familyincluding YOWOv2-Tiny, YOWOv2-Medium, and YOWOv2-Large. We also introduce thepopular dynamic label assignment strategy and anchor-free mechanism to make theYOWOv2 consistent with the advanced model architecture design. With ourimprovement, YOWOv2 is significantly superior to YOWO, and can still keepreal-time detection. Without any bells and whistles, YOWOv2 achieves 87.0 %frame mAP and 52.8 % video mAP with over 20 FPS on the UCF101-24. On the AVA,YOWOv2 achieves 21.7 % frame mAP with over 20 FPS. Our code is available onhttps://github.com/yjh0410/YOWOv2.

YOWOv2: A Stronger yet Efficient Multi-level Detection Framework for Real-time Spatio-temporal Action Detection