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/

Zendy Plus

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

Zendy Tools

Zendy Open

Dynamic scene graph generation (SGG) from videos requires not only acomprehensive understanding of objects across scenes but also a method tocapture the temporal motions and interactions with different objects. Moreover,the long-tailed distribution of visual relationships is a crucial bottleneckfor most dynamic SGG methods. This is because many of them focus on capturingspatio-temporal context using complex architectures, leading to the generationof biased scene graphs. To address these challenges, we propose FloCoDe:Flow-aware Temporal Consistency and Correlation Debiasing with uncertaintyattenuation for unbiased dynamic scene graphs. FloCoDe employs feature warpingusing flow to detect temporally consistent objects across frames. To addressthe long-tail issue of visual relationships, we propose correlation debiasingand a label correlation-based loss to learn unbiased relation representationsfor long-tailed classes. Specifically, we propose to incorporate labelcorrelations using contrastive loss to capture commonly co-occurring relations,which aids in learning robust representations for long-tailed classes. Further,we adopt the uncertainty attenuation-based classifier framework to handle noisyannotations in the SGG data. Extensive experimental evaluation shows aperformance gain as high as 4.1%, demonstrating the superiority of generatingmore unbiased scene graphs.

FloCoDe: Unbiased Dynamic Scene Graph Generation with Temporal Consistency and Correlation Debiasing