by Zoltan Kato, Ting Chuen Pong
Abstract:
A novel video object segmentation method is proposed which aims at combining color and motion information. The model has a multilayer structure: Each feature has its own layer, called feature layer, where a classical Markov random field (MRF) image segmentation model is defined using only the corresponding feature. A special layer is assigned to the combined MRF model, called combined layer, which interacts with each feature layer and provides the segmentation based on the combination of different features. Unlike previous methods, our approach doesnt assume motion boundaries being part of spatial ones. Therefore a very important property of the proposed method is the ability to detect boundaries that are visible only in the motion feature as well as those visible only in the color one. The method is validated on synthetic and real video sequences.
Reference:
Zoltan Kato, Ting Chuen Pong, A Multi-Layer MRF Model for Video Object Segmentation, In Proceedings of Asian Conference on Computer Vision (P. J. Narayanan, Shree K. Nayar, Heung-Yeung Shum, eds.), volume 3852 of Lecture Notes in Computer Science, Hyderabad, India, pp. 953-962, 2006, Springer.
Bibtex Entry:
@string{accv="Proceedings of Asian Conference on Computer Vision"}
@string{lncs="Lecture Notes in Computer Science"}
@string{springer="Springer"}
@InProceedings{Kato-Pong2006,
author = {Kato, Zoltan and Pong, Ting Chuen},
title = {A Multi-Layer {MRF} Model for Video Object
Segmentation},
booktitle = accv,
pages = {953--962},
year = 2006,
editor = {P. J. Narayanan and Shree K. Nayar and Heung-Yeung
Shum},
volume = 3852,
series = lncs,
address = {Hyderabad, India},
month = jan,
publisher = springer,
pdf = {papers/accv2006.pdf},
abstract = {A novel video object segmentation method is proposed
which aims at combining color and motion
information. The model has a multilayer structure:
Each feature has its own layer, called feature
layer, where a classical Markov random field (MRF)
image segmentation model is defined using only the
corresponding feature. A special layer is assigned
to the combined MRF model, called combined layer,
which interacts with each feature layer and provides
the segmentation based on the combination of
different features. Unlike previous methods, our
approach doesnt assume motion boundaries being part
of spatial ones. Therefore a very important property
of the proposed method is the ability to detect
boundaries that are visible only in the motion
feature as well as those visible only in the color
one. The method is validated on synthetic and real
video sequences.}
}