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

Class-incremental semantic image segmentation assumes multiple model updates,each enriching the model to segment new categories. This is typically carriedout by providing expensive pixel-level annotations to the training algorithmfor all new objects, limiting the adoption of such methods in practicalapplications. Approaches that solely require image-level labels offer anattractive alternative, yet, such coarse annotations lack precise informationabout the location and boundary of the new objects. In this paper we arguethat, since classes represent not just indices but semantic entities, theconceptual relationships between them can provide valuable information thatshould be leveraged. We propose a weakly supervised approach that exploits suchsemantic relations to transfer objectness prior from the previously learnedclasses into the new ones, complementing the supervisory signal fromimage-level labels. We validate our approach on a number of continual learningtasks, and show how even a simple pairwise interaction between classes cansignificantly improve the segmentation mask quality of both old and newclasses. We show these conclusions still hold for longer and, hence, morerealistic sequences of tasks and for a challenging few-shot scenario.

RaSP: Relation-aware Semantic Prior for Weakly Supervised Incremental Segmentation