SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2503.10596 · ICCV · 2025

GroundingSuite: Measuring Complex Multi-Granular Pixel Grounding

Wenyu Liu, Xinggang Wang, Yuxuan Zhang, Tianheng Cheng, Xiaoxin Chen, Rui Hu, Lianghui Zhu, Longjin Ran, Lei Liu, Heng Liu

arXiv · PDF · Open in the Atlas

Code that ran

We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. The repositories linked to it are listed below.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Object Multi Object Stuff Class Part Level "The hair of the woman …" "The beak of …" "Paper towels scattered …" "Sidewalk with …" "A cat with green eyes…" RefCOCO Single Object "far left" "in the front" "A trail of white smoke …" "Banana chips …" "A bench in …" gRefCOCO "black car on right and red car" Multi Object "left person" Single Object GranD "a broken computer monitor" Single Object "the water is blue" Stuff Class Single Object Figure 1. Examples of our GroundingSuite dataset. Including four aspects: stuff-class segmentation for context-aware localization, part-level segmentation requiring fine-grained understanding, multi-object segmentation with complex referential relationships, and singleobject segmentation across diverse appearance variations. Note that 'red' color means the referring is wrong.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2503.10596")
get_code_for_paper("2503.10596")
have("2503.10596")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2503.10596.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2503.10596.svg)](https://syntology.ai/paper/2503.10596)

Connect an agent — have() is free.