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

Modern approaches have proved the huge potential of addressing semanticsegmentation as a mask classification task which is widely used ininstance-level segmentation. This paradigm trains models by assigning part ofobject queries to ground truths via conventional one-to-one matching. However,we observe that the popular video semantic segmentation (VSS) dataset haslimited categories per video, meaning less than 10% of queries could be matchedto receive meaningful gradient updates during VSS training. This inefficiencylimits the full expressive potential of all queries.Thus, we present a novelsolution THE-Mask for VSS, which introduces temporal-aware hierarchical objectqueries for the first time. Specifically, we propose to use a simple two-roundmatching mechanism to involve more queries matched with minimal cost duringtraining while without any extra cost during inference. To support ourmore-to-one assignment, in terms of the matching results, we further design ahierarchical loss to train queries with their corresponding hierarchy ofprimary or secondary. Moreover, to effectively capture temporal informationacross frames, we propose a temporal aggregation decoder that fits seamlesslyinto the mask-classification paradigm for VSS. Utilizing temporal-sensitivemulti-level queries, our method achieves state-of-the-art performance on thelatest challenging VSS benchmark VSPW without bells and whistles.

Temporal-aware Hierarchical Mask Classification for Video Semantic Segmentation