Method

Wide Stride Multi-Classification Stereo Matching Network with Subpixel Cross-entropy Loss [WSMCnet]
https://github.com/wyf2017/WSMCnet

Submitted on 20 Jun. 2019 17:19 by
yufeng wang (University of Naval Aviation)

Running time:0.39s
Environment:Nvidia GTX 1070 (Pytorch)

Method Description:
For the stereo matching method based on deep learning, the network architecture is critical for the accuracy of the algorithm, while the efficiency is also an important factor to consider in practical applications. A stereo matching method with spare cost volume in disparity dimension is proposed. The spare cost volume is created by shifting right-view features with a wide stride to reduce greatly the memory and computational resources in three-dimension convolution module. The matching cost is nonlinearly sampled by means of multi-classification in disparity dimension, and model is trained with merging two kind of loss function, so that the accuracy is improved without notably lowering the efficiency.
Parameters:
S=2, C=3, bi-linear
Adam(0.9,0.999),
epoch=20,
lr0=0.001(15)
lr1=0.0001(5)
loss_cross_entry
Latex Bibtex:
@article{wang2019WSMCnet,
title={Stereo Matching Algorithm Based on Three-Dimensional Convolutional Neural Network},
author={Wang, Yufeng and Wang, Hongwei and Yu, Guang and Yang, Mingquan and Yuan, Yuwei and Quan, Jicheng},
journal={Acta Optica Sinica},
volume={39},
number={11},
pages={1115001},
year={2019},
publisher={Chinese Laser Press}
}

Detailed Results

This page provides detailed results for the method(s) selected. For the first 20 test images, the percentage of erroneous pixels is depicted in the table. We use the error metric described in Object Scene Flow for Autonomous Vehicles (CVPR 2015), which considers a pixel to be correctly estimated if the disparity or flow end-point error is <3px or <5% (for scene flow this criterion needs to be fulfilled for both disparity maps and the flow map). Underneath, the left input image, the estimated results and the error maps are shown (for disp_0/disp_1/flow/scene_flow, respectively). The error map uses the log-color scale described in Object Scene Flow for Autonomous Vehicles (CVPR 2015), depicting correct estimates (<3px or <5% error) in blue and wrong estimates in red color tones. Dark regions in the error images denote the occluded pixels which fall outside the image boundaries. The false color maps of the results are scaled to the largest ground truth disparity values / flow magnitudes.

Test Set Average

Error D1-bg D1-fg D1-all
All / All 1.72 4.19 2.13
All / Est 1.72 4.19 2.13
Noc / All 1.51 3.57 1.85
Noc / Est 1.51 3.57 1.85
This table as LaTeX

Test Image 0

Error D1-bg D1-fg D1-all
All / All 1.62 1.08 1.55
All / Est 1.62 1.08 1.55
Noc / All 1.55 1.08 1.48
Noc / Est 1.55 1.08 1.48
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 1

Error D1-bg D1-fg D1-all
All / All 1.57 1.64 1.57
All / Est 1.57 1.64 1.57
Noc / All 1.42 1.64 1.45
Noc / Est 1.42 1.64 1.45
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 2

Error D1-bg D1-fg D1-all
All / All 2.40 1.49 2.35
All / Est 2.40 1.49 2.35
Noc / All 2.22 1.49 2.18
Noc / Est 2.22 1.49 2.18
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 3

Error D1-bg D1-fg D1-all
All / All 2.17 0.04 1.98
All / Est 2.17 0.04 1.98
Noc / All 2.05 0.04 1.87
Noc / Est 2.05 0.04 1.87
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 4

Error D1-bg D1-fg D1-all
All / All 0.92 0.31 0.82
All / Est 0.92 0.31 0.82
Noc / All 0.84 0.31 0.76
Noc / Est 0.84 0.31 0.76
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 5

Error D1-bg D1-fg D1-all
All / All 3.49 0.96 3.26
All / Est 3.49 0.96 3.26
Noc / All 2.81 0.96 2.64
Noc / Est 2.81 0.96 2.64
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 6

Error D1-bg D1-fg D1-all
All / All 4.21 0.28 3.80
All / Est 4.21 0.28 3.80
Noc / All 4.25 0.28 3.82
Noc / Est 4.25 0.28 3.82
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 7

Error D1-bg D1-fg D1-all
All / All 0.47 2.93 0.95
All / Est 0.47 2.93 0.95
Noc / All 0.44 2.93 0.93
Noc / Est 0.44 2.93 0.93
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 8

Error D1-bg D1-fg D1-all
All / All 0.41 1.47 0.61
All / Est 0.41 1.47 0.61
Noc / All 0.41 1.47 0.60
Noc / Est 0.41 1.47 0.60
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 9

Error D1-bg D1-fg D1-all
All / All 0.47 1.14 0.64
All / Est 0.47 1.14 0.64
Noc / All 0.46 1.14 0.63
Noc / Est 0.46 1.14 0.63
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 10

Error D1-bg D1-fg D1-all
All / All 1.14 2.05 1.35
All / Est 1.14 2.05 1.35
Noc / All 1.12 2.05 1.33
Noc / Est 1.12 2.05 1.33
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 11

Error D1-bg D1-fg D1-all
All / All 0.97 0.36 0.86
All / Est 0.97 0.36 0.86
Noc / All 0.95 0.36 0.84
Noc / Est 0.95 0.36 0.84
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 12

Error D1-bg D1-fg D1-all
All / All 0.69 0.81 0.70
All / Est 0.69 0.81 0.70
Noc / All 0.57 0.81 0.58
Noc / Est 0.57 0.81 0.58
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 13

Error D1-bg D1-fg D1-all
All / All 0.68 0.31 0.64
All / Est 0.68 0.31 0.64
Noc / All 0.52 0.31 0.50
Noc / Est 0.52 0.31 0.50
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 14

Error D1-bg D1-fg D1-all
All / All 1.45 0.00 1.43
All / Est 1.45 0.00 1.43
Noc / All 1.32 0.00 1.30
Noc / Est 1.32 0.00 1.30
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 15

Error D1-bg D1-fg D1-all
All / All 2.85 0.50 2.64
All / Est 2.85 0.50 2.64
Noc / All 2.83 0.50 2.62
Noc / Est 2.83 0.50 2.62
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 16

Error D1-bg D1-fg D1-all
All / All 3.79 0.17 3.26
All / Est 3.79 0.17 3.26
Noc / All 3.57 0.17 3.07
Noc / Est 3.57 0.17 3.07
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 17

Error D1-bg D1-fg D1-all
All / All 0.81 0.23 0.75
All / Est 0.81 0.23 0.75
Noc / All 0.77 0.23 0.72
Noc / Est 0.77 0.23 0.72
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 18

Error D1-bg D1-fg D1-all
All / All 5.16 1.71 3.53
All / Est 5.16 1.71 3.53
Noc / All 5.01 1.71 3.43
Noc / Est 5.01 1.71 3.43
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 19

Error D1-bg D1-fg D1-all
All / All 0.83 0.45 0.79
All / Est 0.83 0.45 0.79
Noc / All 0.75 0.45 0.72
Noc / Est 0.75 0.45 0.72
This table as LaTeX

Input Image

D1 Result

D1 Error




eXTReMe Tracker