Haochen-Wang409/TreeVGR
[ICLR'26] Traceable Evidence Enhanced Visual Grounded Reasoning: Evaluation and Methodology
This project offers a specialized benchmark called TreeBench and a model called TreeVGR for evaluating and improving how AI models 'think with images.' It takes complex images, particularly those with many objects, and outputs an assessment of an AI's ability to answer detailed questions by pointing to specific visual evidence within the image. It is designed for AI researchers and developers who are building and testing advanced visual reasoning AI.
Use this if you are developing or evaluating AI models that need to precisely locate and reason about multiple objects and their interactions within complex images.
Not ideal if you are looking for a general-purpose image recognition tool or a model that doesn't require explicit traceability of visual evidence.
Stars
77
Forks
2
Language
Python
License
Apache-2.0
Category
Last pushed
Jan 26, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/Haochen-Wang409/TreeVGR"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
col14m/cadrille
[ICLR2026] cadrille: Multi-modal CAD Reconstruction with Online Reinforcement Learning
filaPro/cad-recode
[ICCV2025] CAD-Recode: Reverse Engineering CAD Code from Point Clouds
pengsongyou/openscene
[CVPR'23] OpenScene: 3D Scene Understanding with Open Vocabularies
worldbench/3EED
[NeurIPS 2025 DB Track] 3EED: Ground Everything Everywhere in 3D
cambrian-mllm/cambrian-s
Cambrian-S: Towards Spatial Supersensing in Video