Method

Accurate and Efficient Stereo Matching via Attention Concatenation Volume [Fast-ACVNet+]
https://github.com/gangweiX/Fast-ACVNet

Submitted on 8 Aug. 2024 05:04 by
Gangwei Xu (Huazhong University of Science and Technology)

Running time:0.05 s
Environment:NVIDIA RTX 3090 (PyTorch)

Method Description:
We design a fast version of ACV to enable real-
time performance, named Fast-ACV+, which generates
high likelihood disparity hypotheses and the
corresponding attention weights from correlation
clues to significantly reduce computational and
memory costs and meanwhile maintain a satisfactory
accuracy.
Parameters:
We finetune the pre-trained Scene Flow model on the
mixed KITTI 2012 and KITTI 2015 training sets for
500 epochs.
Latex Bibtex:
@article{xu2023accurate,
title={Accurate and efficient stereo matching
via attention concatenation volume},
author={Xu, Gangwei and Wang, Yun and Cheng,
Junda and Tang, Jinhui and Yang, Xin},
journal={IEEE Transactions on Pattern Analysis
and Machine Intelligence},
year={2023},
publisher={IEEE}
}

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 1.70 3.53 2.01
All / Est 1.70 3.53 2.01
Noc / All 1.56 3.29 1.85
Noc / Est 1.56 3.29 1.85
This table as LaTeX

Test Image 0

Error D1-bg D1-fg D1-all
All / All 1.69 2.06 1.74
All / Est 1.69 2.06 1.74
Noc / All 1.64 2.06 1.70
Noc / Est 1.64 2.06 1.70
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 1

Error D1-bg D1-fg D1-all
All / All 2.15 3.50 2.31
All / Est 2.15 3.50 2.31
Noc / All 2.10 3.50 2.25
Noc / Est 2.10 3.50 2.25
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 2

Error D1-bg D1-fg D1-all
All / All 2.18 8.46 2.48
All / Est 2.18 8.46 2.48
Noc / All 1.93 8.46 2.26
Noc / Est 1.93 8.46 2.26
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 3

Error D1-bg D1-fg D1-all
All / All 2.48 0.53 2.30
All / Est 2.48 0.53 2.30
Noc / All 2.36 0.53 2.19
Noc / Est 2.36 0.53 2.19
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 4

Error D1-bg D1-fg D1-all
All / All 0.66 0.23 0.59
All / Est 0.66 0.23 0.59
Noc / All 0.65 0.23 0.58
Noc / Est 0.65 0.23 0.58
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 5

Error D1-bg D1-fg D1-all
All / All 2.90 3.02 2.91
All / Est 2.90 3.02 2.91
Noc / All 2.71 3.02 2.74
Noc / Est 2.71 3.02 2.74
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 6

Error D1-bg D1-fg D1-all
All / All 4.81 1.70 4.48
All / Est 4.81 1.70 4.48
Noc / All 4.91 1.70 4.57
Noc / Est 4.91 1.70 4.57
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 7

Error D1-bg D1-fg D1-all
All / All 0.39 8.80 2.04
All / Est 0.39 8.80 2.04
Noc / All 0.40 8.80 2.07
Noc / Est 0.40 8.80 2.07
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 8

Error D1-bg D1-fg D1-all
All / All 0.40 1.94 0.68
All / Est 0.40 1.94 0.68
Noc / All 0.39 1.94 0.68
Noc / Est 0.39 1.94 0.68
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 9

Error D1-bg D1-fg D1-all
All / All 0.45 1.95 0.83
All / Est 0.45 1.95 0.83
Noc / All 0.45 2.05 0.85
Noc / Est 0.45 2.05 0.85
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 10

Error D1-bg D1-fg D1-all
All / All 1.14 2.52 1.45
All / Est 1.14 2.52 1.45
Noc / All 1.14 2.52 1.46
Noc / Est 1.14 2.52 1.46
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 11

Error D1-bg D1-fg D1-all
All / All 0.75 0.65 0.73
All / Est 0.75 0.65 0.73
Noc / All 0.76 0.65 0.74
Noc / Est 0.76 0.65 0.74
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 12

Error D1-bg D1-fg D1-all
All / All 0.77 1.33 0.81
All / Est 0.77 1.33 0.81
Noc / All 0.62 1.33 0.67
Noc / Est 0.62 1.33 0.67
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 13

Error D1-bg D1-fg D1-all
All / All 0.64 0.13 0.58
All / Est 0.64 0.13 0.58
Noc / All 0.62 0.13 0.56
Noc / Est 0.62 0.13 0.56
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 14

Error D1-bg D1-fg D1-all
All / All 1.38 0.00 1.36
All / Est 1.38 0.00 1.36
Noc / All 1.26 0.00 1.24
Noc / Est 1.26 0.00 1.24
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 15

Error D1-bg D1-fg D1-all
All / All 2.86 0.19 2.62
All / Est 2.86 0.19 2.62
Noc / All 2.92 0.19 2.66
Noc / Est 2.92 0.19 2.66
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 16

Error D1-bg D1-fg D1-all
All / All 3.90 0.05 3.34
All / Est 3.90 0.05 3.34
Noc / All 3.69 0.05 3.15
Noc / Est 3.69 0.05 3.15
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 17

Error D1-bg D1-fg D1-all
All / All 0.98 0.18 0.90
All / Est 0.98 0.18 0.90
Noc / All 0.97 0.18 0.89
Noc / Est 0.97 0.18 0.89
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 18

Error D1-bg D1-fg D1-all
All / All 4.66 1.84 3.32
All / Est 4.66 1.84 3.32
Noc / All 4.64 1.84 3.30
Noc / Est 4.64 1.84 3.30
This table as LaTeX

Input Image

D1 Result

D1 Error


Test Image 19

Error D1-bg D1-fg D1-all
All / All 0.92 0.82 0.91
All / Est 0.92 0.82 0.91
Noc / All 0.92 0.82 0.91
Noc / Est 0.92 0.82 0.91
This table as LaTeX

Input Image

D1 Result

D1 Error




eXTReMe Tracker