PointSIFT: A SIFT-like Network Module for 3D Point Cloud Semantic Segmentation (2018-07-02T00:00:00.000000Z)

TL;DR

Inspired by the outstanding 2D shape descriptor SIFT, a module called PointSIFT is designed that encodes information of different orientations and is adaptive to scale of shape, which outperforms state-of-the-art method on standard benchmark datasets.

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.

Authors

Cewu Lu

2 papers

Mingyang Jiang

1 papers

Yiran Wu

1 papers

TL;DR

Abstract

Authors

References38 items

PointCNN

Dynamic Graph CNN for Learning on Point Clouds

Large-Scale Point Cloud Semantic Segmentation with Superpoint Graphs

Frustum PointNets for 3D Object Detection from RGB-D Data

3D object classification and retrieval with Spherical CNNs

SEGCloud: Semantic Segmentation of 3D Point Clouds

2D-Driven 3D Object Detection in RGB-D Images

SceneCut: Joint Geometric and Object Segmentation for Indoor Scenes

PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space

Octree Generating Networks: Efficient Convolutional Architectures for High-resolution 3D Outputs

ScanNet: Richly-Annotated 3D Reconstructions of Indoor Scenes

PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation

SyncSpecCNN: Synchronized Spectral CNN for 3D Shape Segmentation

OctNet: Learning Deep 3D Representations at High Resolutions

3D Semantic Parsing of Large-Scale Indoor Spaces

FPNN: Field Probing Neural Networks for 3D Data

Cut Pursuit: Fast Algorithms to Learn Piecewise Constant Functions

Volumetric and Multi-view CNNs for Object Classification on 3D Data

G-CNN: An Iterative Grid Based Object Detector

Deep Residual Learning for Image Recognition

3D-Assisted Feature Synthesis for Novel Views of an Object

Deep Sliding Shapes for Amodal 3D Object Detection in RGB-D Images

SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation

VoxNet: A 3D Convolutional Neural Network for real-time object recognition

Multi-view Convolutional Neural Networks for 3D Shape Recognition

Geodesic Convolutional Neural Networks on Riemannian Manifolds

Fully convolutional networks for semantic segmentation

Meta-representation of shape families

3D ShapeNets: A deep representation for volumetric shapes

Spectral Networks and Locally Connected Networks on Graphs

Consistent Shape Maps via Semidefinite Programming

Voxel Cloud Connectivity Segmentation - Supervoxels for Point Clouds

Distinctive Image Features from Scale-Invariant Keypoints

Large-scale Point Cloud Semantic Segmentation with Superpoint Graphs ( SUPPLEMENTARY ) 1 . Model Details

Notes on the OpenSURF Library

ArXiv eprints

Distinctive Image Features from Scale-Invariant Keypoints

Tchapmi , Christopher B . Choy , Iro Armeni , JunYoung Gwak , and Silvio Savarese . Segcloud : Semantic segmentation of 3 d point clouds

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names