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

Unsupervised object discovery (UOD) refers to the task of discriminating thewhole region of objects from the background within a scene without relying onlabeled datasets, which benefits the task of bounding-box-level localizationand pixel-level segmentation. This task is promising due to its ability todiscover objects in a generic manner. We roughly categorise existing techniquesinto two main directions, namely the generative solutions based on imageresynthesis, and the clustering methods based on self-supervised models. Wehave observed that the former heavily relies on the quality of imagereconstruction, while the latter shows limitations in effectively modelingsemantic correlations. To directly target at object discovery, we focus on thelatter approach and propose a novel solution by incorporating weakly-supervisedcontrastive learning (WCL) to enhance semantic information exploration. Wedesign a semantic-guided self-supervised learning model to extract high-levelsemantic features from images, which is achieved by fine-tuning the featureencoder of a self-supervised model, namely DINO, via WCL. Subsequently, weintroduce Principal Component Analysis (PCA) to localize object regions. Theprincipal projection direction, corresponding to the maximal eigenvalue, servesas an indicator of the object region(s). Extensive experiments on benchmarkunsupervised object discovery datasets demonstrate the effectiveness of ourproposed solution. The source code and experimental results are publiclyavailable via our project page at https://github.com/npucvr/WSCUOD.git.

Weakly-supervised Contrastive Learning for Unsupervised Object Discovery