PASS3D: Precise and Accelerated Semantic Segmentation for 3D Point Cloud

TL;DR

PASS3D is a fast and accurate 3D point cloud segmentation framework that combines geometric and deep learning methods, improving efficiency and robustness for autonomous driving applications.

Contribution

It introduces a two-stage approach with an accelerated clustering algorithm and a neural network for semantic labeling, along with a novel data augmentation technique.

Findings

Outperforms state-of-the-art on KITTI dataset

Achieves high recall with fewer proposals

Demonstrates real-time capability in autonomous driving scenarios

Abstract

In this paper, we propose PASS3D to achieve point-wise semantic segmentation for 3D point cloud. Our framework combines the efficiency of traditional geometric methods with robustness of deep learning methods, consisting of two stages: At stage-1, our accelerated cluster proposal algorithm will generate refined cluster proposals by segmenting point clouds without ground, capable of generating less redundant proposals with higher recall in an extremely short time; stage-2 we will amplify and further process these proposals by a neural network to estimate semantic label for each point and meanwhile propose a novel data augmentation method to enhance the network's recognition capability for all categories especially for non-rigid objects. Evaluated on KITTI raw dataset, PASS3D stands out against the state-of-the-art on some results, making itself competent to 3D perception in autonomous driving system. Our source code will be open-sourced. A video demonstration is available at https://www.youtube.com/watch?v=cukEqDuP_Qw.