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

In this technical report, we introduce our solution to human-centricspatio-temporal video grounding task. We propose a concise and effectiveframework named STVGFormer, which models spatiotemporal visual-linguisticdependencies with a static branch and a dynamic branch. The static branchperforms cross-modal understanding in a single frame and learns to localize thetarget object spatially according to intra-frame visual cues like objectappearances. The dynamic branch performs cross-modal understanding acrossmultiple frames. It learns to predict the starting and ending time of thetarget moment according to dynamic visual cues like motions. Both the staticand dynamic branches are designed as cross-modal transformers. We furtherdesign a novel static-dynamic interaction block to enable the static anddynamic branches to transfer useful and complementary information from eachother, which is shown to be effective to improve the prediction on hard cases.Our proposed method achieved 39.6% vIoU and won the first place in the HC-STVGtrack of the 4th Person in Context Challenge.

STVGFormer: Spatio-Temporal Video Grounding with Static-Dynamic Cross-Modal Understanding