Benchmark RadarCOCOTree
COCOTree evaluates open tree-structured visual decomposition, segmenting images into hierarchical trees of visual components with unconstrained granularity. It includes over 21K images and 1.8M structural nodes, with 3.5K unique labels, and uses the Open Tree Quality (OTQ) metric for scoring.
- Released
- 2026-05-21
- Readiness
- Runnable
- Primary field
- General AI
Why it matters
Provides a standardized evaluation for a new task paradigm, enabling comparison across models on open vocabulary and long-tail visual structures, where existing benchmarks lack hierarchical decomposition metrics.
Motivation
We formalize and enable the task of open tree decomposition, which segments an image into hierarchical trees of visual components with unconstrained granularity and flexibility.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.