PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation

TL;DR

PointSIFT introduces an orientation-encoding module inspired by 2D SIFT to enhance 3D point cloud feature extraction, improving semantic segmentation performance across standard benchmarks.

Contribution

The paper proposes PointSIFT, a novel module for 3D point cloud analysis that encodes orientation and scale, integrated into PointNet architectures for better shape representation.

Findings

Outperforms state-of-the-art methods on benchmark datasets

Enhances feature representation in 3D point cloud segmentation

Demonstrates effectiveness of orientation and multi-scale encoding

Abstract

Recently, 3D understanding research sheds light on extracting features from point cloud directly, which requires effective shape pattern description of point clouds. Inspired by the outstanding 2D shape descriptor SIFT, we design a module called PointSIFT that encodes information of different orientations and is adaptive to scale of shape. Specifically, an orientation-encoding unit is designed to describe eight crucial orientations, and multi-scale representation is achieved by stacking several orientation-encoding units. PointSIFT module can be integrated into various PointNet-based architecture to improve the representation ability. Extensive experiments show our PointSIFT-based framework outperforms state-of-the-art method on standard benchmark datasets. The code and trained model will be published accompanied by this paper.