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/

Zendy Plus

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

Zendy Tools

Zendy Open

Recognizing instances at different scales simultaneously is a fundamentalchallenge in visual detection problems. While spatial multi-scale modeling hasbeen well studied in object detection, how to effectively apply a multi-scalearchitecture to temporal models for activity detection is still under-explored.In this paper, we identify three unique challenges that need to be specificallyhandled for temporal activity detection compared to its spatial counterpart. Toaddress all these issues, we propose Dynamic Temporal Pyramid Network (DTPN), anew activity detection framework with a multi-scale pyramidal architecturefeaturing three novel designs: (1) We sample input video frames dynamicallywith varying frame per seconds (FPS) to construct a natural pyramidal input forvideo of an arbitrary length. (2) We design a two-branch multi-scale temporalfeature hierarchy to deal with the inherent temporal scale variation ofactivity instances. (3) We further exploit the temporal context of activitiesby appropriately fusing multi-scale feature maps, and demonstrate that bothlocal and global temporal contexts are important. By combining all thesecomponents into a uniform network, we end up with a single-shot activitydetector involving single-pass inferencing and end-to-end training. Extensiveexperiments show that the proposed DTPN achieves state-of-the-art performanceon the challenging ActvityNet dataset.

Dynamic Temporal Pyramid Network: A Closer Look at Multi-Scale Modeling for Activity Detection