
LoSA, a memory-and-parameter-efficient backbone adapter, enables end-to-end training of large video foundation models for improved temporal action localization in untrimmed videos.


coremsg

scaling-end-to-end-temporal-action-localization-with-long-short-range-adapters


Scaling End-to-End Temporal Action Localization with Long-Short-range Adapters


title_rewrite


A multi-level cross-scale solution called video self-stitching graph network (VSGN) is proposed to tackle the challenge of large action scale variation, especially for short actions, in temporal action localization.


video-self-stitching-graph-network-for-improving-temporal-action-localization-of-short-actions


Video Self-Stitching Graph Network for Improving Temporal Action Localization of Short Actions
