We present a supervised learning-based method to estimate a per-pixel
confidence for optical flow vectors. Regions of low texture and pixels
close to occlusion boundaries are known to be difficult for optical
flow algorithms. Using a spatiotemporal feature vector, we estimate
if a flow algorithm is likely to fail in a given region. Our method
is not restricted to any specific class of flow algorithm and does
not make any scene specific assumptions. By automatically learning
this confidence, we can combine the output of several computed flow
fields from different algorithms to select the best performing algorithm
per pixel. Our optical flow confidence measure allows one to achieve
better overall results by discarding the most troublesome pixels.
We illustrate the effectiveness of our method on four different optical
flow algorithms over a variety of real and synthetic sequences. For
algorithm selection, we achieve the top overall results on a large
test set, and at times even surpass the results of the best algorithm
among the candidates.
%0 Journal Article
%1 MacAodha2013
%A Mac Aodha, Oisin
%A Humayun, Ahmad
%A Pollefeys, Marc
%A Brostow, Gabriel J
%D 2013
%J IEEE TPAMI
%K Accuracy,Adaptive Forest,Supervised algorithms,Random data flow,Optical imaging,Optical learning,Vectors,algorithm measure,synthetic measurement,Prediction optics,Optical selection,confidence variables
%N 5
%P 1107--20
%R 10.1109/TPAMI.2012.171
%T Learning a confidence measure for optical flow.
%U http://dl.acm.org/citation.cfm?id=2478536.2478616
%V 35
%X We present a supervised learning-based method to estimate a per-pixel
confidence for optical flow vectors. Regions of low texture and pixels
close to occlusion boundaries are known to be difficult for optical
flow algorithms. Using a spatiotemporal feature vector, we estimate
if a flow algorithm is likely to fail in a given region. Our method
is not restricted to any specific class of flow algorithm and does
not make any scene specific assumptions. By automatically learning
this confidence, we can combine the output of several computed flow
fields from different algorithms to select the best performing algorithm
per pixel. Our optical flow confidence measure allows one to achieve
better overall results by discarding the most troublesome pixels.
We illustrate the effectiveness of our method on four different optical
flow algorithms over a variety of real and synthetic sequences. For
algorithm selection, we achieve the top overall results on a large
test set, and at times even surpass the results of the best algorithm
among the candidates.
@article{MacAodha2013,
abstract = {We present a supervised learning-based method to estimate a per-pixel
confidence for optical flow vectors. Regions of low texture and pixels
close to occlusion boundaries are known to be difficult for optical
flow algorithms. Using a spatiotemporal feature vector, we estimate
if a flow algorithm is likely to fail in a given region. Our method
is not restricted to any specific class of flow algorithm and does
not make any scene specific assumptions. By automatically learning
this confidence, we can combine the output of several computed flow
fields from different algorithms to select the best performing algorithm
per pixel. Our optical flow confidence measure allows one to achieve
better overall results by discarding the most troublesome pixels.
We illustrate the effectiveness of our method on four different optical
flow algorithms over a variety of real and synthetic sequences. For
algorithm selection, we achieve the top overall results on a large
test set, and at times even surpass the results of the best algorithm
among the candidates.},
added-at = {2013-09-29T14:16:50.000+0200},
author = {{Mac Aodha}, Oisin and Humayun, Ahmad and Pollefeys, Marc and Brostow, Gabriel J},
biburl = {https://www.bibsonomy.org/bibtex/22c793037dd8bccaeee534c91ee76554b/guillem.palou},
doi = {10.1109/TPAMI.2012.171},
interhash = {9fa8d946e405f680bc65cf2c3840f04e},
intrahash = {2c793037dd8bccaeee534c91ee76554b},
issn = {1939-3539},
journal = {IEEE TPAMI},
keywords = {Accuracy,Adaptive Forest,Supervised algorithms,Random data flow,Optical imaging,Optical learning,Vectors,algorithm measure,synthetic measurement,Prediction optics,Optical selection,confidence variables},
number = 5,
pages = {1107--20},
pmid = {22868652},
timestamp = {2013-09-29T14:16:50.000+0200},
title = {{Learning a confidence measure for optical flow.}},
url = {http://dl.acm.org/citation.cfm?id=2478536.2478616},
volume = 35,
year = 2013
}