SegAgent: Exploring Pixel Understanding Capabilities in MLLMs by Imitating Human Annotator Trajectories
Paper
•
2503.08625
•
Published
•
26
Guangkai Xu
guangkaixu
AI & ML interests
Monocular Depth Estimation, 3D Reconstruction, Visual-Language Models
Recent Activity
upvoted
a
paper
about 1 month ago
SegAgent: Exploring Pixel Understanding Capabilities in MLLMs by
Imitating Human Annotator Trajectories
liked
a dataset
6 months ago
sayakpaul/nyu_depth_v2
updated
a Space
6 months ago
guangkaixu/GenPercept
Organizations
