DailyPapers (@HuggingPapers)

2026-01-31 | โค๏ธ 98 | ๐Ÿ” 10


LingBot-Depth: Masked Depth Modeling for Spatial Perception

Transforms incomplete, noisy depth sensor data into high-quality metric-accurate 3D measurements via masked depth modeling. Trained on 3M RGB-depth pairs, it outperforms top-tier RGB-D cameras and serves as a spatial perception foundation for robotics.

๋ฏธ๋””์–ด

image


Tags

3D-Vision rendering