Pyramid Stereo Matching Network (2018-03-23T00:00:00.000000Z)

TL;DR

PSMNet is a pyramid stereo matching network consisting of two main modules: spatial pyramid pooling and 3D CNN, which takes advantage of the capacity of global context information by aggregating context in different scales and locations to form a cost volume.

Abstract

Recent work has shown that depth estimation from a stereo pair of images can be formulated as a supervised learning task to be resolved with convolutional neural networks (CNNs). However, current architectures rely on patch-based Siamese networks, lacking the means to exploit context information for finding correspondence in ill-posed regions. To tackle this problem, we propose PSMNet, a pyramid stereo matching network consisting of two main modules: spatial pyramid pooling and 3D CNN. The spatial pyramid pooling module takes advantage of the capacity of global context information by aggregating context in different scales and locations to form a cost volume. The 3D CNN learns to regularize cost volume using stacked multiple hourglass networks in conjunction with intermediate supervision. The proposed approach was evaluated on several benchmark datasets. Our method ranked first in the KITTI 2012 and 2015 leaderboards before March 18, 2018. The codes of PSMNet are available at: https://github.com/JiaRenChang/PSMNet.

Authors

Jia-Ren Chang

2 papers

Yong-Sheng Chen

2 papers

TL;DR

Abstract

Authors

References33 items

Learning for Disparity Estimation Through Feature Constancy

Learning Deep Correspondence through Prior and Posterior Feature Constancy

PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume

Cascade Residual Learning: A Two-Stage Convolutional Neural Network for Stereo Matching

Stacked Deconvolutional Network for Semantic Segmentation

SGM-Nets: Semi-Global Matching with Neural Networks

Rethinking Atrous Convolution for Semantic Image Segmentation

End-to-End Learning of Geometry and Context for Deep Stereo Regression

Label Refinement Network for Coarse-to-Fine Semantic Segmentation

Improved Stereo Matching with Constant Highway Networks and Reflective Confidence Learning

Detect, Replace, Refine: Deep Structured Prediction for Pixel Wise Labeling

Pyramid Scene Parsing Network

RefineNet: Multi-path Refinement Networks for High-Resolution Semantic Segmentation

Optical Flow Estimation Using a Spatial Pyramid Network

Efficient Deep Learning for Stereo Matching

DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs

Learning to Refine Object Segments

Stacked Hourglass Networks for Human Pose Estimation

Deep Residual Learning for Image Recognition

A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation

3D Object Proposals for Accurate Object Class Detection

MeshStereo: A Global Stereo Model with Mesh Alignment Regularization for View Interpolation

Multi-Scale Context Aggregation by Dilated Convolutions

Stereo Matching by Training a Convolutional Neural Network to Compare Image Patches

ParseNet: Looking Wider to See Better

Displets: Resolving stereo ambiguities using object knowledge

U-Net: Convolutional Networks for Biomedical Image Segmentation

Fast R-CNN

Fully convolutional networks for semantic segmentation

Neural Machine Translation by Jointly Learning to Align and Translate

Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition

A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms

Accurate and Efficient Stereo Processing by Semi-Global Matching and Mutual Information

Field of Study

Journal Information

Name

Page