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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Video Moment Retrieval (VMR) aims to retrieve temporal segments in untrimmedvideos corresponding to a given language query by constructing cross-modalalignment strategies. However, these existing strategies are often sub-optimalsince they ignore the modality imbalance problem, \textit{i.e.}, the semanticrichness inherent in videos far exceeds that of a given limited-lengthsentence. Therefore, in pursuit of better alignment, a natural idea isenhancing the video modality to filter out query-irrelevant semantics, andenhancing the text modality to capture more segment-relevant knowledge. In thispaper, we introduce Modal-Enhanced Semantic Modeling (MESM), a novel frameworkfor more balanced alignment through enhancing features at two levels. First, weenhance the video modality at the frame-word level through word reconstruction.This strategy emphasizes the portions associated with query words inframe-level features while suppressing irrelevant parts. Therefore, theenhanced video contains less redundant semantics and is more balanced with thetextual modality. Second, we enhance the textual modality at thesegment-sentence level by learning complementary knowledge from contextsentences and ground-truth segments. With the knowledge added to the query, thetextual modality thus maintains more meaningful semantics and is more balancedwith the video modality. By implementing two levels of MESM, the semanticinformation from both modalities is more balanced to align, thereby bridgingthe modality gap. Experiments on three widely used benchmarks, including theout-of-distribution settings, show that the proposed framework achieves a newstart-of-the-art performance with notable generalization ability (e.g., 4.42%and 7.69% average gains of R1@0.7 on Charades-STA and Charades-CG). The codewill be available at https://github.com/lntzm/MESM.

Towards Balanced Alignment: Modal-Enhanced Semantic Modeling for Video Moment Retrieval