.png){kind=icon}
.png){kind=icon}
Return
KD-DiffSeg: knowledge distillation guided LiDAR-camera diffusion framework for 3D semantic segmentation
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2026.132272.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We enhances 3D segmentation by combining diffusion models and knowledge distillation. • We aligns image features to point cloud features through diffusion model. • We enables targeted knowledge transfer without relying on large teacher models. • We achieves superior performance on SemanticKITTI and nuScenes with low parameters.
Keywords:
3D semantic segmentation
knowledge distillation
diffusion models
LiDAR-camera fusion
point cloud features
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W