PANet: LiDAR Panoptic Segmentation with Sparse Instance Proposal and Aggregation

TL;DR

PANet introduces a novel LiDAR panoptic segmentation framework that employs a sparse instance proposal and aggregation method, eliminating the need for offset branches and enhancing large object segmentation accuracy.

Contribution

The paper presents a non-learning sparse instance proposal module and an instance aggregation technique, improving large object segmentation without relying on offset prediction.

Findings

Achieves state-of-the-art results on SemanticKITTI and nuScenes datasets.

Effectively improves large object segmentation accuracy.

Eliminates dependency on offset branches in LiDAR panoptic segmentation.

Abstract

Reliable LiDAR panoptic segmentation (LPS), including both semantic and instance segmentation, is vital for many robotic applications, such as autonomous driving. This work proposes a new LPS framework named PANet to eliminate the dependency on the offset branch and improve the performance on large objects, which are always over-segmented by clustering algorithms. Firstly, we propose a non-learning Sparse Instance Proposal (SIP) module with the ``sampling-shifting-grouping" scheme to directly group thing points into instances from the raw point cloud efficiently. More specifically, balanced point sampling is introduced to generate sparse seed points with more uniform point distribution over the distance range. And a shift module, termed bubble shifting, is proposed to shrink the seed points to the clustered centers. Then we utilize the connected component label algorithm to generate instance proposals. Furthermore, an instance aggregation module is devised to integrate potentially fragmented instances, improving the performance of the SIP module on large objects. Extensive experiments show that PANet achieves state-of-the-art performance among published works on the SemanticKITII validation and nuScenes validation for the panoptic segmentation task.