Method

DeepPruner: Learning Efficient Stereo Matching via Differentiable PatchMatch [DeepPruner (best)]


Submitted on 23 Jun. 2019 05:44 by
Shivam Duggal (Uber)

Running time:0.18 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
Our goal is to significantly speed up the runtime of
current state-of-the-art stereo algorithms to enable
real-time inference. Towards this goal, we developed a
differentiable PatchMatch module that allows us to
discard most disparities without requiring full cost
volume evaluation. We then exploit this representation
to learn which range to prune for each pixel. By
progressively reducing the search space and
effectively propagating such information, we are able
to efficiently compute the cost volume for high
likelihood hypotheses and achieve savings in both
memory and computation. Finally, an image guided
refinement module is exploited to further improve the
performance. Since all our components are
differentiable, the full network can be trained end-to-
end. Our experiments show that our method achieves
competitive results on KITTI and SceneFlow datasets
while running in real-time at 62ms.
Parameters:
Please refer to the paper:
https://arxiv.org/abs/1909.05845
Latex Bibtex:
@inproceedings{Duggal2019ICCV,
title = {DeepPruner: Learning Efficient Stereo Matching
via Differentiable PatchMatch},
author = {Shivam Duggal and Shenlong Wang and Wei-
Chiu Ma and Rui Hu and Raquel Urtasun},
booktitle = {ICCV},
year = {2019}}

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 1.87 3.56 2.15
All / Est 1.87 3.56 2.15
Noc / All 1.71 3.18 1.95
Noc / Est 1.71 3.18 1.95
This table as LaTeX

Test Image 0

Error D1-bg D1-fg D1-all
All / All 1.76 0.73 1.62
All / Est 1.76 0.73 1.62
Noc / All 1.73 0.73 1.59
Noc / Est 1.73 0.73 1.59
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 1

Error D1-bg D1-fg D1-all
All / All 1.99 3.10 2.12
All / Est 1.99 3.10 2.12
Noc / All 1.92 3.10 2.05
Noc / Est 1.92 3.10 2.05
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 2

Error D1-bg D1-fg D1-all
All / All 3.72 3.94 3.73
All / Est 3.72 3.94 3.73
Noc / All 3.56 3.94 3.58
Noc / Est 3.56 3.94 3.58
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 3

Error D1-bg D1-fg D1-all
All / All 3.24 0.57 2.99
All / Est 3.24 0.57 2.99
Noc / All 3.23 0.57 2.98
Noc / Est 3.23 0.57 2.98
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 4

Error D1-bg D1-fg D1-all
All / All 0.78 0.35 0.71
All / Est 0.78 0.35 0.71
Noc / All 0.75 0.35 0.68
Noc / Est 0.75 0.35 0.68
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 5

Error D1-bg D1-fg D1-all
All / All 2.83 2.73 2.82
All / Est 2.83 2.73 2.82
Noc / All 2.53 2.73 2.54
Noc / Est 2.53 2.73 2.54
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 6

Error D1-bg D1-fg D1-all
All / All 5.12 1.70 4.76
All / Est 5.12 1.70 4.76
Noc / All 5.23 1.70 4.85
Noc / Est 5.23 1.70 4.85
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 7

Error D1-bg D1-fg D1-all
All / All 0.50 4.57 1.30
All / Est 0.50 4.57 1.30
Noc / All 0.51 4.57 1.32
Noc / Est 0.51 4.57 1.32
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 8

Error D1-bg D1-fg D1-all
All / All 0.59 4.03 1.22
All / Est 0.59 4.03 1.22
Noc / All 0.58 4.03 1.22
Noc / Est 0.58 4.03 1.22
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 9

Error D1-bg D1-fg D1-all
All / All 0.68 2.69 1.19
All / Est 0.68 2.69 1.19
Noc / All 0.69 2.82 1.22
Noc / Est 0.69 2.82 1.22
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 10

Error D1-bg D1-fg D1-all
All / All 1.24 2.41 1.51
All / Est 1.24 2.41 1.51
Noc / All 1.26 2.41 1.52
Noc / Est 1.26 2.41 1.52
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 11

Error D1-bg D1-fg D1-all
All / All 0.83 0.80 0.83
All / Est 0.83 0.80 0.83
Noc / All 0.83 0.80 0.83
Noc / Est 0.83 0.80 0.83
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 12

Error D1-bg D1-fg D1-all
All / All 0.83 1.12 0.85
All / Est 0.83 1.12 0.85
Noc / All 0.68 1.12 0.71
Noc / Est 0.68 1.12 0.71
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 13

Error D1-bg D1-fg D1-all
All / All 0.94 0.00 0.82
All / Est 0.94 0.00 0.82
Noc / All 0.87 0.00 0.76
Noc / Est 0.87 0.00 0.76
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 14

Error D1-bg D1-fg D1-all
All / All 1.29 0.17 1.28
All / Est 1.29 0.17 1.28
Noc / All 1.20 0.17 1.18
Noc / Est 1.20 0.17 1.18
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 15

Error D1-bg D1-fg D1-all
All / All 3.05 0.11 2.78
All / Est 3.05 0.11 2.78
Noc / All 3.10 0.11 2.83
Noc / Est 3.10 0.11 2.83
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 16

Error D1-bg D1-fg D1-all
All / All 4.20 0.18 3.61
All / Est 4.20 0.18 3.61
Noc / All 4.00 0.18 3.43
Noc / Est 4.00 0.18 3.43
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 17

Error D1-bg D1-fg D1-all
All / All 0.86 0.23 0.80
All / Est 0.86 0.23 0.80
Noc / All 0.84 0.23 0.77
Noc / Est 0.84 0.23 0.77
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 18

Error D1-bg D1-fg D1-all
All / All 5.48 2.09 3.87
All / Est 5.48 2.09 3.87
Noc / All 5.40 2.09 3.81
Noc / Est 5.40 2.09 3.81
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 19

Error D1-bg D1-fg D1-all
All / All 1.15 1.25 1.16
All / Est 1.15 1.25 1.16
Noc / All 1.15 1.25 1.16
Noc / Est 1.15 1.25 1.16
This table as LaTeX

Input Image

D1 Result

D1 Error




eXTReMe Tracker