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

Panoptic segmentation is posed as a new popular test-bed for thestate-of-the-art holistic scene understanding methods with the requirement ofsimultaneously segmenting both foreground things and background stuff. Thestate-of-the-art panoptic segmentation network exhibits high structuralcomplexity in different network components, i.e. backbone, proposal-basedforeground branch, segmentation-based background branch, and feature fusionmodule across branches, which heavily relies on expert knowledge and tedioustrials. In this work, we propose an efficient, cooperative and highly automatedframework to simultaneously search for all main components including backbone,segmentation branches, and feature fusion module in a unified panopticsegmentation pipeline based on the prevailing one-shot Network ArchitectureSearch (NAS) paradigm. Notably, we extend the common single-task NAS into themulti-component scenario by taking the advantage of the newly proposedintra-modular search space and problem-oriented inter-modular search space,which helps us to obtain an optimal network architecture that not only performswell in both instance segmentation and semantic segmentation tasks but also beaware of the reciprocal relations between foreground things and backgroundstuff classes. To relieve the vast computation burden incurred by applying NASto complicated network architectures, we present a novel path-priority greedysearch policy to find a robust, transferrable architecture with significantlyreduced searching overhead. Our searched architecture, namely Auto-Panoptic,achieves the new state-of-the-art on the challenging COCO and ADE20Kbenchmarks. Moreover, extensive experiments are conducted to demonstrate theeffectiveness of path-priority policy and transferability of Auto-Panopticacross different datasets. Codes and models are available at:https://github.com/Jacobew/AutoPanoptic.

Auto-Panoptic: Cooperative Multi-Component Architecture Search for Panoptic Segmentation