Method

Automatic Labeling Environment (Joint Segmentation and Stereo) [ALE-Stereo]
http://www.inf.ethz.ch/personal/ladickyl/

Submitted on 13 Dec. 2013 09:23 by
Andreas Geiger (MPI Tübingen)

Running time:50 min
Environment:1 core @ 3.0 Ghz (C/C++)

Method Description:
Joint Optimisation for Object Class Segmentation
and Dense Stereo Reconstruction

The problems of dense stereo reconstruction and
object class segmentation can both be formulated
as Conditional Random Field based labelling
problems, in which every pixel in the image is
assigned a label corresponding to either its
disparity, or an object class such as road or
building. While these two problems are mutually
informative, no attempt has been made to jointly
optimise their labellings. In this work we
provide a principled energy minimisation
framework that unifies the two problems and
demonstrate that, by resolving ambiguities in
real world data, joint optimisation of the two
problems substantially improves performance.
Parameters:
The default parameter set has been used. The stereo
parameters have been set according to the KITTI
calibration files with a maximum disparity value of
150. The system has been trained with the semantic
KITTI labels provided at
http://www.inf.ethz.ch/personal/ladickyl/.
Latex Bibtex:
@INPROCEEDINGS{Ladicky2010BMVC,
author = {Ladicky, Lubor and Sturgess, Paul and
Russell, Christopher and Sengupta, Sunando and
Bastanlar, Yalin and Clocksin, William and Torr,
Philip},
title = {Joint Optimisation for Object Class
Segmentation and Dense Stereo Reconstruction},
booktitle = {BMVC},
year = {2010}
}

Detailed Results

This page provides detailed results for the method(s) selected. For each of the first 20 test images, the number of erroneous pixels at all thresholds is depicted in the table. Underneath, the left input image, the disparity / end-point error map and the estimated (and interpolated) disparity / optical flow map are shown. The error map scales linearly between 0 (black) and >=5 (white) pixels error. Red denotes all occluded pixels, falling outside the image boundaries. The false color map is scaled to the largest ground truth disparity / flow value.

Test Set Average

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 59.87 % 60.48 % 13.0 px 13.5 px
3 pixels 50.48 % 51.19 % 13.0 px 13.5 px
4 pixels 44.68 % 45.43 % 13.0 px 13.5 px
5 pixels 40.80 % 41.55 % 13.0 px 13.5 px
This table as LaTeX

Reflective Regions

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 87.62 % 88.07 % 24.6 px 25.4 px
3 pixels 83.80 % 84.37 % 24.6 px 25.4 px
4 pixels 80.56 % 81.22 % 24.6 px 25.4 px
5 pixels 77.60 % 78.34 % 24.6 px 25.4 px
This table as LaTeX

Test Image 0

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 55.79 % 56.65 % 8.7 px 9.2 px
3 pixels 44.91 % 45.86 % 8.7 px 9.2 px
4 pixels 37.43 % 38.50 % 8.7 px 9.2 px
5 pixels 31.77 % 32.84 % 8.7 px 9.2 px
This table as LaTeX





Test Image 1

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 70.20 % 70.58 % 17.6 px 18.7 px
3 pixels 58.32 % 59.07 % 17.6 px 18.7 px
4 pixels 48.80 % 49.93 % 17.6 px 18.7 px
5 pixels 43.30 % 44.65 % 17.6 px 18.7 px
This table as LaTeX





Test Image 2

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 61.35 % 61.83 % 7.9 px 8.0 px
3 pixels 49.55 % 50.25 % 7.9 px 8.0 px
4 pixels 42.30 % 43.12 % 7.9 px 8.0 px
5 pixels 37.82 % 38.72 % 7.9 px 8.0 px
This table as LaTeX





Test Image 3

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 41.87 % 42.77 % 5.1 px 5.2 px
3 pixels 32.07 % 32.98 % 5.1 px 5.2 px
4 pixels 27.02 % 27.86 % 5.1 px 5.2 px
5 pixels 24.47 % 25.29 % 5.1 px 5.2 px
This table as LaTeX





Test Image 4

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 49.08 % 50.15 % 16.3 px 17.4 px
3 pixels 40.90 % 42.15 % 16.3 px 17.4 px
4 pixels 37.04 % 38.38 % 16.3 px 17.4 px
5 pixels 34.69 % 36.08 % 16.3 px 17.4 px
This table as LaTeX





Test Image 5

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 40.34 % 40.33 % 6.6 px 6.6 px
3 pixels 31.46 % 31.33 % 6.6 px 6.6 px
4 pixels 28.33 % 28.22 % 6.6 px 6.6 px
5 pixels 26.20 % 26.14 % 6.6 px 6.6 px
This table as LaTeX





Test Image 6

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 59.08 % 59.61 % 11.9 px 12.3 px
3 pixels 50.94 % 51.64 % 11.9 px 12.3 px
4 pixels 47.88 % 48.67 % 11.9 px 12.3 px
5 pixels 45.78 % 46.62 % 11.9 px 12.3 px
This table as LaTeX





Test Image 7

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 49.20 % 50.37 % 14.5 px 14.8 px
3 pixels 41.90 % 43.13 % 14.5 px 14.8 px
4 pixels 38.85 % 40.04 % 14.5 px 14.8 px
5 pixels 36.66 % 37.74 % 14.5 px 14.8 px
This table as LaTeX





Test Image 8

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 63.60 % 63.31 % 11.6 px 11.5 px
3 pixels 51.26 % 51.05 % 11.6 px 11.5 px
4 pixels 42.84 % 42.70 % 11.6 px 11.5 px
5 pixels 38.49 % 38.36 % 11.6 px 11.5 px
This table as LaTeX





Test Image 9

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 46.26 % 47.42 % 8.8 px 8.9 px
3 pixels 37.79 % 38.70 % 8.8 px 8.9 px
4 pixels 34.93 % 35.72 % 8.8 px 8.9 px
5 pixels 33.14 % 33.68 % 8.8 px 8.9 px
This table as LaTeX





Test Image 10

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 59.66 % 60.04 % 12.1 px 12.4 px
3 pixels 52.38 % 52.73 % 12.1 px 12.4 px
4 pixels 46.44 % 46.75 % 12.1 px 12.4 px
5 pixels 41.43 % 41.71 % 12.1 px 12.4 px
This table as LaTeX





Test Image 11

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 63.70 % 64.91 % 12.1 px 12.7 px
3 pixels 55.24 % 56.70 % 12.1 px 12.7 px
4 pixels 51.07 % 52.62 % 12.1 px 12.7 px
5 pixels 48.41 % 49.99 % 12.1 px 12.7 px
This table as LaTeX





Test Image 12

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 49.24 % 50.55 % 6.6 px 7.9 px
3 pixels 35.64 % 37.30 % 6.6 px 7.9 px
4 pixels 29.40 % 31.22 % 6.6 px 7.9 px
5 pixels 27.52 % 29.38 % 6.6 px 7.9 px
This table as LaTeX





Test Image 13

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 74.38 % 74.70 % 12.0 px 12.2 px
3 pixels 64.87 % 65.28 % 12.0 px 12.2 px
4 pixels 57.16 % 57.64 % 12.0 px 12.2 px
5 pixels 50.51 % 51.04 % 12.0 px 12.2 px
This table as LaTeX





Test Image 14

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 62.41 % 62.56 % 9.6 px 9.7 px
3 pixels 52.47 % 52.58 % 9.6 px 9.7 px
4 pixels 46.87 % 46.94 % 9.6 px 9.7 px
5 pixels 42.60 % 42.66 % 9.6 px 9.7 px
This table as LaTeX





Test Image 15

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 71.14 % 71.85 % 33.4 px 34.5 px
3 pixels 66.33 % 67.14 % 33.4 px 34.5 px
4 pixels 63.89 % 64.75 % 33.4 px 34.5 px
5 pixels 61.78 % 62.69 % 33.4 px 34.5 px
This table as LaTeX





Test Image 16

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 78.02 % 78.39 % 28.3 px 28.1 px
3 pixels 69.45 % 69.72 % 28.3 px 28.1 px
4 pixels 61.39 % 61.43 % 28.3 px 28.1 px
5 pixels 54.74 % 54.39 % 28.3 px 28.1 px
This table as LaTeX





Test Image 17

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 48.43 % 49.10 % 14.1 px 14.4 px
3 pixels 43.77 % 44.48 % 14.1 px 14.4 px
4 pixels 41.78 % 42.06 % 14.1 px 14.4 px
5 pixels 40.00 % 39.86 % 14.1 px 14.4 px
This table as LaTeX





Test Image 18

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 82.92 % 83.14 % 11.8 px 11.9 px
3 pixels 75.50 % 75.81 % 11.8 px 11.9 px
4 pixels 67.68 % 68.07 % 11.8 px 11.9 px
5 pixels 60.03 % 60.53 % 11.8 px 11.9 px
This table as LaTeX





Test Image 19

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 50.08 % 51.07 % 6.7 px 6.8 px
3 pixels 38.91 % 40.23 % 6.7 px 6.8 px
4 pixels 31.89 % 33.40 % 6.7 px 6.8 px
5 pixels 28.13 % 29.74 % 6.7 px 6.8 px
This table as LaTeX







eXTReMe Tracker