Real-time estimation of optical flow based on optimized haar wavelet features
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Standard
Real-time estimation of optical flow based on optimized haar wavelet features. / Salmen, Jan; Caup, Lukas; Igel, Christian.
Evolutionary Multi-Criterion Optimization: 6th International Conference, EMO 2011, Ouro Preto, Brazil, April 5-8, 2011. Proceedings. ed. / Ricardo H. C. Takahashi; Kalyanmoy Deb; Elizabeth F. Wanner; Salvatore Greco. Springer, 2011. p. 448-461 (Lecture notes in computer science, Vol. 6576).Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - Real-time estimation of optical flow based on optimized haar wavelet features
AU - Salmen, Jan
AU - Caup, Lukas
AU - Igel, Christian
PY - 2011
Y1 - 2011
N2 - Estimation of optical flow is required in many computer vision applications.These applications often have to deal with strict time constraints. Therefore,flow algorithms with both high accuracy and computational efficiency aredesirable. Accordingly, designing such a flow algorithm involves multi-objectiveoptimization. In this work, we build on a popular algorithm developed for realtime applications. It is originally based on the Census transform and benefitsfrom this encoding for table-based matching and tracking of interest points. Wepropose to use the more universal Haar wavelet features instead of the Censustransform within the same framework. The resulting approach is more flexible,in particular it allows for sub-pixel accuracy. For comparison with the originalmethod and another baseline algorithm, we considered both popular benchmarkdatasets as well as a long synthetic video sequence. We employed evolutionarymulti-objective optimization to tune the algorithms. This allows to compare thedifferent approaches in a systematic and unbiased way. Our results show thatthe overall performance of our method is significantly higher compared to thereference implementation.
AB - Estimation of optical flow is required in many computer vision applications.These applications often have to deal with strict time constraints. Therefore,flow algorithms with both high accuracy and computational efficiency aredesirable. Accordingly, designing such a flow algorithm involves multi-objectiveoptimization. In this work, we build on a popular algorithm developed for realtime applications. It is originally based on the Census transform and benefitsfrom this encoding for table-based matching and tracking of interest points. Wepropose to use the more universal Haar wavelet features instead of the Censustransform within the same framework. The resulting approach is more flexible,in particular it allows for sub-pixel accuracy. For comparison with the originalmethod and another baseline algorithm, we considered both popular benchmarkdatasets as well as a long synthetic video sequence. We employed evolutionarymulti-objective optimization to tune the algorithms. This allows to compare thedifferent approaches in a systematic and unbiased way. Our results show thatthe overall performance of our method is significantly higher compared to thereference implementation.
U2 - 10.1007/978-3-642-19893-9_31
DO - 10.1007/978-3-642-19893-9_31
M3 - Article in proceedings
SN - 978-3-642-19892-2
T3 - Lecture notes in computer science
SP - 448
EP - 461
BT - Evolutionary Multi-Criterion Optimization
A2 - Takahashi, Ricardo H. C.
A2 - Deb, Kalyanmoy
A2 - Wanner, Elizabeth F.
A2 - Greco, Salvatore
PB - Springer
T2 - Evolutionary Multi-Criterion Optimization
Y2 - 5 April 2011 through 8 April 2011
ER -
ID: 168459867