Method

Appearance-Aligned Block-Matching Stereo [AABM]


Submitted on 17 Nov. 2014 17:11 by
Nils Einecke (Honda Research Institute Europe GmbH)

Running time:0.12 s
Environment:1 core @ 3.1 Ghz (C/C++)

Method Description:
You can get a copy of the paper here.

Appearance-aligned stereo warps one stereo image according to the expected road geometry for reducing the perspective appearance change between the left and right stereo image. The resulting disparity map is combined with the disparity map of stereo computation without warped images by means of a pixel-wise correlation value comparison. This way warping model coherent pixels are mainly drawn from the aligned disparity map and model incoherent pixels are drawn from the normal disparity map.

For the actual stereo computation a standard block-matching stereo with SNCC cost function is used. The images are pre-processed with a Gaussian filter and resized to half the image size. After the stereo processing the disparity maps are rescaled to full image size. Furthermore, standard post-processed is applied, including a left-right check, error island removal (region grouping), hole-filling and median filtering.

Please note that the current results are updated using a newer hardware and a new software version with some bugs fixes. The original results were from 04.01.2013.

Timings: block-matching SNCC stereo (75ms), LR-check (18ms), region growing (2ms), hole-filling (2ms), median filter (23ms)

Runtime includes some IPP routines but no multi-core or GPU usage.
Parameters:
linear warping model: x2 = x + a*y + n
a=0.25, n=-21.6
SNCC (first stage 3x3, second stage 9x9)
sigma_gauss = 1.0
disp_range = 128
min correlation threshold = 0.3
region growing threshold = 2.5 disparity
min region size = 200 pixel
median filter = 1x9 and 9x1
Latex Bibtex:
@InProceedings{Einecke13,
author = {N. Einecke and J. Eggert},
title = {Stereo Image Warping for Improved Depth Estimation of Road Surfaces},
booktitle = {IV},
pages = {89-194},
year = 2013
}

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 7.43 % 8.77 % 1.0 px 1.2 px
3 pixels 4.97 % 6.04 % 1.0 px 1.2 px
4 pixels 3.79 % 4.65 % 1.0 px 1.2 px
5 pixels 3.06 % 3.77 % 1.0 px 1.2 px
This table as LaTeX

Reflective Regions

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 31.34 % 34.25 % 3.8 px 4.4 px
3 pixels 23.33 % 26.25 % 3.8 px 4.4 px
4 pixels 18.65 % 21.45 % 3.8 px 4.4 px
5 pixels 15.59 % 18.22 % 3.8 px 4.4 px
This table as LaTeX

Test Image 0

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 8.35 % 10.68 % 1.0 px 1.1 px
3 pixels 5.89 % 7.79 % 1.0 px 1.1 px
4 pixels 4.48 % 6.00 % 1.0 px 1.1 px
5 pixels 3.47 % 4.65 % 1.0 px 1.1 px
This table as LaTeX





Test Image 1

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 7.17 % 9.41 % 1.0 px 1.1 px
3 pixels 4.62 % 6.26 % 1.0 px 1.1 px
4 pixels 3.53 % 4.63 % 1.0 px 1.1 px
5 pixels 2.86 % 3.60 % 1.0 px 1.1 px
This table as LaTeX





Test Image 2

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 6.97 % 7.52 % 0.8 px 0.8 px
3 pixels 4.17 % 4.41 % 0.8 px 0.8 px
4 pixels 2.87 % 3.03 % 0.8 px 0.8 px
5 pixels 2.10 % 2.21 % 0.8 px 0.8 px
This table as LaTeX





Test Image 3

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 8.89 % 10.96 % 0.8 px 1.0 px
3 pixels 5.61 % 7.34 % 0.8 px 1.0 px
4 pixels 3.52 % 4.90 % 0.8 px 1.0 px
5 pixels 2.33 % 3.32 % 0.8 px 1.0 px
This table as LaTeX





Test Image 4

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 8.11 % 9.00 % 0.8 px 0.9 px
3 pixels 6.04 % 6.10 % 0.8 px 0.9 px
4 pixels 4.14 % 4.22 % 0.8 px 0.9 px
5 pixels 1.94 % 2.06 % 0.8 px 0.9 px
This table as LaTeX





Test Image 5

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 3.22 % 3.54 % 0.6 px 0.6 px
3 pixels 1.80 % 1.89 % 0.6 px 0.6 px
4 pixels 1.31 % 1.32 % 0.6 px 0.6 px
5 pixels 1.04 % 1.05 % 0.6 px 0.6 px
This table as LaTeX





Test Image 6

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 11.16 % 12.14 % 1.4 px 1.5 px
3 pixels 7.09 % 7.84 % 1.4 px 1.5 px
4 pixels 5.52 % 6.05 % 1.4 px 1.5 px
5 pixels 4.78 % 5.25 % 1.4 px 1.5 px
This table as LaTeX





Test Image 7

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 9.76 % 11.68 % 1.6 px 1.7 px
3 pixels 6.09 % 7.75 % 1.6 px 1.7 px
4 pixels 4.60 % 5.91 % 1.6 px 1.7 px
5 pixels 3.73 % 4.71 % 1.6 px 1.7 px
This table as LaTeX





Test Image 8

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 3.71 % 4.28 % 0.6 px 0.7 px
3 pixels 2.05 % 2.48 % 0.6 px 0.7 px
4 pixels 1.46 % 1.90 % 0.6 px 0.7 px
5 pixels 1.20 % 1.64 % 0.6 px 0.7 px
This table as LaTeX





Test Image 9

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 14.68 % 16.79 % 1.3 px 1.6 px
3 pixels 8.55 % 10.81 % 1.3 px 1.6 px
4 pixels 5.50 % 7.73 % 1.3 px 1.6 px
5 pixels 4.09 % 5.99 % 1.3 px 1.6 px
This table as LaTeX





Test Image 10

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 4.25 % 5.55 % 0.7 px 0.7 px
3 pixels 2.75 % 3.60 % 0.7 px 0.7 px
4 pixels 2.04 % 2.63 % 0.7 px 0.7 px
5 pixels 1.59 % 1.94 % 0.7 px 0.7 px
This table as LaTeX





Test Image 11

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 14.63 % 17.07 % 1.5 px 1.7 px
3 pixels 9.82 % 11.84 % 1.5 px 1.7 px
4 pixels 7.19 % 8.83 % 1.5 px 1.7 px
5 pixels 5.91 % 7.16 % 1.5 px 1.7 px
This table as LaTeX





Test Image 12

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 7.43 % 9.70 % 1.5 px 1.6 px
3 pixels 5.87 % 7.99 % 1.5 px 1.6 px
4 pixels 5.06 % 6.73 % 1.5 px 1.6 px
5 pixels 4.73 % 6.05 % 1.5 px 1.6 px
This table as LaTeX





Test Image 13

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 8.47 % 9.63 % 1.3 px 1.4 px
3 pixels 5.89 % 6.99 % 1.3 px 1.4 px
4 pixels 4.95 % 5.80 % 1.3 px 1.4 px
5 pixels 4.25 % 5.08 % 1.3 px 1.4 px
This table as LaTeX





Test Image 14

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 3.78 % 4.29 % 0.6 px 0.7 px
3 pixels 2.21 % 2.50 % 0.6 px 0.7 px
4 pixels 1.75 % 1.99 % 0.6 px 0.7 px
5 pixels 1.52 % 1.74 % 0.6 px 0.7 px
This table as LaTeX





Test Image 15

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 7.84 % 8.34 % 0.9 px 0.9 px
3 pixels 4.78 % 4.90 % 0.9 px 0.9 px
4 pixels 3.12 % 3.06 % 0.9 px 0.9 px
5 pixels 2.37 % 2.31 % 0.9 px 0.9 px
This table as LaTeX





Test Image 16

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 3.87 % 4.48 % 0.6 px 0.7 px
3 pixels 1.61 % 2.12 % 0.6 px 0.7 px
4 pixels 0.82 % 1.09 % 0.6 px 0.7 px
5 pixels 0.50 % 0.61 % 0.6 px 0.7 px
This table as LaTeX





Test Image 17

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 8.56 % 9.54 % 1.1 px 1.2 px
3 pixels 6.37 % 7.00 % 1.1 px 1.2 px
4 pixels 4.73 % 5.09 % 1.1 px 1.2 px
5 pixels 3.64 % 3.87 % 1.1 px 1.2 px
This table as LaTeX





Test Image 18

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 7.61 % 8.31 % 0.8 px 0.8 px
3 pixels 5.82 % 6.36 % 0.8 px 0.8 px
4 pixels 4.82 % 5.28 % 0.8 px 0.8 px
5 pixels 4.08 % 4.48 % 0.8 px 0.8 px
This table as LaTeX





Test Image 19

Error Out-Noc Out-All Avg-Noc Avg-All
2 pixels 3.47 % 4.20 % 0.7 px 0.7 px
3 pixels 1.98 % 2.16 % 0.7 px 0.7 px
4 pixels 1.53 % 1.51 % 0.7 px 0.7 px
5 pixels 1.31 % 1.27 % 0.7 px 0.7 px
This table as LaTeX







eXTReMe Tracker