Felix Heide (@FelixHeide)
2026-01-02 | โค๏ธ 278 | ๐ 38
Starting the new year without human labeling ๐!! Multimodal lidar-camera data is a gold mine of dense 3D geometry hiding in plain sight. For supervised pretraining and validation at scale at @torc_robotics, we rely on fully automated pseudo-labeling pipelines. Exploiting geometric priors from temporally accumulated LiDAR maps and an iterative update rule enforces joint geometricโsemantic consistency while detecting moving objects via inconsistencies.
We achieve 3D semantic labels and 3D bounding boxes with human-like quality at 200m+ range required for highway driving.
Paper: https://light.princeton.edu/publication/unilips/
Exciting work with @torc_robotics with Filippo Ghilotti, Samuel Brucker, Nahku Saidy, Matteo Matteucci, Mario Bijelic.
๐ ์๋ณธ ๋งํฌ
๋ฏธ๋์ด
๐ Related
Auto-generated - needs manual review