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

Scene, as the crucial unit of storytelling in movies, contains complexactivities of actors and their interactions in a physical environment.Identifying the composition of scenes serves as a critical step towardssemantic understanding of movies. This is very challenging -- compared to thevideos studied in conventional vision problems, e.g. action recognition, asscenes in movies usually contain much richer temporal structures and morecomplex semantic information. Towards this goal, we scale up the scenesegmentation task by building a large-scale video dataset MovieScenes, whichcontains 21K annotated scene segments from 150 movies. We further propose alocal-to-global scene segmentation framework, which integrates multi-modalinformation across three levels, i.e. clip, segment, and movie. This frameworkis able to distill complex semantics from hierarchical temporal structures overa long movie, providing top-down guidance for scene segmentation. Ourexperiments show that the proposed network is able to segment a movie intoscenes with high accuracy, consistently outperforming previous methods. We alsofound that pretraining on our MovieScenes can bring significant improvements tothe existing approaches.

A Local-to-Global Approach to Multi-modal Movie Scene Segmentation