Discover, Segment, and Select: A Progressive Mechanism for Zero-shot Camouflaged Object Segmentation

Yilong Yang, Jianxin Tian, Shengchuan Zhang, Liujuan Cao · 2026

Read Discover, Segment, and Select: A Progressive Mechanism for Zero-shot Camouflaged Object Segmentation by Yilong Yang, Jianxin Tian, Shengchuan Zhang, Liujuan Cao (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-19944.txt.

Description

Camouflaged animals and objects blend into their surroundings, so finding them in an image and outlining them precisely are two different hard problems. A 2026 arXiv paper by Yilong Yang and three co-authors proposes a three-step mechanism, Discover-Segment-Select, for zero-shot camouflaged object segmentation.

Current pipelines usually ask a multimodal model to find the object, then hand the result to a segmentation model. The first step often misplaces objects, invents extras, or misses some entirely. The new method first generates diverse object proposals from visual features. It then refines each through segmentation. Finally a multimodal model judges the candidate masks and picks the best. With no training or supervision, it leads prior methods on several benchmarks, above all in scenes with multiple instances.

Catalog record

CREATORYilong Yang, Jianxin Tian, Shengchuan Zhang, Liujuan Cao
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-19944

Browse the full library · About the Tower of Babel