Method

Deep Discrete Flow [DDF]
https://ps.is.tuebingen.mpg.de/publications/guney2016accv

Submitted on 2 Sep. 2016 17:04 by
Fatma Guney (MPI)

Running time:~1 min
Environment:GPU @ 2.5 Ghz (C/C++)

Method Description:
Motivated by the success of deep learning techniques in matching problems, we
present a method for learning context-aware features for solving optical flow
using discrete optimization. Towards this goal, we present an efficient way of
training a context network with a large receptive field size on top of a local
network using dilated convolutions on patches. We perform feature matching by
comparing each pixel in the reference image to every pixel in the target image,
utilizing fast GPU matrix multiplication. The matching cost volume from the
network’s output forms the data term for discrete MAP inference in a pairwise
Markov random field.
Parameters:
See supplementary.
Latex Bibtex:
@conference{Guney2016ACCV,
title = {Deep Discrete Flow},
author = {G{\"u}ney, Fatma and Geiger, Andreas},
booktitle = {Asian Conference on Computer Vision
(ACCV)},
year = {2016}
}

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 Fl-bg Fl-fg Fl-all
All / All 20.36 25.19 21.17
All / Est 20.36 25.19 21.17
Noc / All 10.44 21.32 12.41
Noc / Est 10.44 21.32 12.41
This table as LaTeX

Test Image 0

Error Fl-bg Fl-fg Fl-all
All / All 7.57 29.95 10.64
All / Est 7.57 29.95 10.64
Noc / All 4.12 29.95 8.04
Noc / Est 4.12 29.95 8.04
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 1

Error Fl-bg Fl-fg Fl-all
All / All 7.91 37.35 11.20
All / Est 7.91 37.35 11.20
Noc / All 3.07 37.35 7.30
Noc / Est 3.07 37.35 7.30
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 2

Error Fl-bg Fl-fg Fl-all
All / All 19.11 20.58 19.18
All / Est 19.11 20.58 19.18
Noc / All 8.98 20.58 9.65
Noc / Est 8.98 20.58 9.65
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 3

Error Fl-bg Fl-fg Fl-all
All / All 23.11 96.52 29.89
All / Est 23.11 96.52 29.89
Noc / All 10.77 95.36 18.06
Noc / Est 10.77 95.36 18.06
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 4

Error Fl-bg Fl-fg Fl-all
All / All 29.68 59.48 34.62
All / Est 29.68 59.48 34.62
Noc / All 18.20 58.05 25.65
Noc / Est 18.20 58.05 25.65
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 5

Error Fl-bg Fl-fg Fl-all
All / All 19.56 0.00 17.81
All / Est 19.56 0.00 17.81
Noc / All 12.56 0.00 11.27
Noc / Est 12.56 0.00 11.27
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 6

Error Fl-bg Fl-fg Fl-all
All / All 17.60 1.53 15.91
All / Est 17.60 1.53 15.91
Noc / All 11.85 1.53 10.62
Noc / Est 11.85 1.53 10.62
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 7

Error Fl-bg Fl-fg Fl-all
All / All 0.45 100.00 19.94
All / Est 0.45 100.00 19.94
Noc / All 0.45 100.00 19.07
Noc / Est 0.45 100.00 19.07
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 8

Error Fl-bg Fl-fg Fl-all
All / All 0.62 12.32 2.78
All / Est 0.62 12.32 2.78
Noc / All 0.62 12.32 2.78
Noc / Est 0.62 12.32 2.78
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 9

Error Fl-bg Fl-fg Fl-all
All / All 0.30 71.94 18.60
All / Est 0.30 71.94 18.60
Noc / All 0.30 71.94 18.60
Noc / Est 0.30 71.94 18.60
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 10

Error Fl-bg Fl-fg Fl-all
All / All 10.77 10.50 10.71
All / Est 10.77 10.50 10.71
Noc / All 4.04 10.50 5.72
Noc / Est 4.04 10.50 5.72
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 11

Error Fl-bg Fl-fg Fl-all
All / All 11.47 9.40 11.10
All / Est 11.47 9.40 11.10
Noc / All 4.69 9.41 5.64
Noc / Est 4.69 9.41 5.64
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 12

Error Fl-bg Fl-fg Fl-all
All / All 14.93 6.22 14.35
All / Est 14.93 6.22 14.35
Noc / All 4.06 6.22 4.24
Noc / Est 4.06 6.22 4.24
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 13

Error Fl-bg Fl-fg Fl-all
All / All 14.36 27.23 15.94
All / Est 14.36 27.23 15.94
Noc / All 2.31 10.87 3.31
Noc / Est 2.31 10.87 3.31
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 14

Error Fl-bg Fl-fg Fl-all
All / All 14.87 14.45 14.86
All / Est 14.87 14.45 14.86
Noc / All 3.02 14.45 3.26
Noc / Est 3.02 14.45 3.26
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 15

Error Fl-bg Fl-fg Fl-all
All / All 30.66 9.70 28.76
All / Est 30.66 9.70 28.76
Noc / All 13.00 9.70 12.62
Noc / Est 13.00 9.70 12.62
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 16

Error Fl-bg Fl-fg Fl-all
All / All 32.79 38.50 33.63
All / Est 32.79 38.50 33.63
Noc / All 15.90 38.50 19.96
Noc / Est 15.90 38.50 19.96
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 17

Error Fl-bg Fl-fg Fl-all
All / All 23.19 24.57 23.34
All / Est 23.19 24.57 23.34
Noc / All 6.89 24.57 9.18
Noc / Est 6.89 24.57 9.18
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 18

Error Fl-bg Fl-fg Fl-all
All / All 39.32 100.00 68.15
All / Est 39.32 100.00 68.15
Noc / All 20.45 100.00 42.39
Noc / Est 20.45 100.00 42.39
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 19

Error Fl-bg Fl-fg Fl-all
All / All 34.54 16.96 32.55
All / Est 34.54 16.96 32.55
Noc / All 16.60 16.96 16.65
Noc / Est 16.60 16.96 16.65
This table as LaTeX

Input Image

Flow Result

Flow Error




eXTReMe Tracker