.png){kind=icon}
.png){kind=icon}
Return
SMFNet: Stacking multi-frame network for 4D spatial-temporal LiDAR semantic segmentation
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2025.132233.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• This work provides a novel idea to process a 4D semantic segmentation task of LiDAR in a 3D way, which saves computational and memory costs and accelerates the training process. • This work proposes a spatio-temporal framework, consisting of a Spatial-Aware Feature Learning (SAFL) module and a Temporal-Aware Feature Learning (TAFL) module, to extract spatial and temporal information in a unified pattern. • This work designs a Motion-Aware Feature Learning module (MAFL), which processes a residual image with a specifically designed 2D network, to strengthen the moving feature learning and facilitate the prediction of moving objects.