Variational Inference for Visual Tracking

Jaco Vermaak, Neil D. Lawrence, Patrick Pérez
,  I:773-780, 2003.

Abstract

The likelihood models used in probabilistic visual tracking applications are often complex non-linear and/or non-Gaussian functions, leading to analytically intractable inference. Solutions then require numerical approximation techniques, of which the particle filter is a popular choice. Particle filters, however, degrade in performance as the dimensionality of the state space increases and the support of the likelihood decreases. As an alternative to particle filters this paper introduces a variational approximation to the tracking recursion. The variational inference is intractable in itself, and is combined with an efficient importance sampling procedure to obtain the required estimates. The algorithm is shown to compare favourably with particle filtering techniques on a synthetic example and two real tracking problems. The first involves the tracking of a designated object in a video sequence based on its colour properties, whereas the second involves contour extraction in a single image.

Cite this Paper


BibTeX
@InProceedings{pmlr-v-vermaak-variational03, title = {Variational Inference for Visual Tracking}, author = {Jaco Vermaak and Neil D. Lawrence and Patrick Pérez}, pages = {773--780}, year = {}, editor = {}, volume = {I}, url = {http://inverseprobability.com/publications/vermaak-variational03.html}, abstract = {The likelihood models used in probabilistic visual tracking applications are often complex non-linear and/or non-Gaussian functions, leading to analytically intractable inference. Solutions then require numerical approximation techniques, of which the particle filter is a popular choice. Particle filters, however, degrade in performance as the dimensionality of the state space increases and the support of the likelihood decreases. As an alternative to particle filters this paper introduces a variational approximation to the tracking recursion. The variational inference is intractable in itself, and is combined with an efficient importance sampling procedure to obtain the required estimates. The algorithm is shown to compare favourably with particle filtering techniques on a synthetic example and two real tracking problems. The first involves the tracking of a designated object in a video sequence based on its colour properties, whereas the second involves contour extraction in a single image.} }
Endnote
%0 Conference Paper %T Variational Inference for Visual Tracking %A Jaco Vermaak %A Neil D. Lawrence %A Patrick Pérez %B %C Proceedings of Machine Learning Research %D %E %F pmlr-v-vermaak-variational03 %I PMLR %J Proceedings of Machine Learning Research %P 773--780 %U http://inverseprobability.com %V %W PMLR %X The likelihood models used in probabilistic visual tracking applications are often complex non-linear and/or non-Gaussian functions, leading to analytically intractable inference. Solutions then require numerical approximation techniques, of which the particle filter is a popular choice. Particle filters, however, degrade in performance as the dimensionality of the state space increases and the support of the likelihood decreases. As an alternative to particle filters this paper introduces a variational approximation to the tracking recursion. The variational inference is intractable in itself, and is combined with an efficient importance sampling procedure to obtain the required estimates. The algorithm is shown to compare favourably with particle filtering techniques on a synthetic example and two real tracking problems. The first involves the tracking of a designated object in a video sequence based on its colour properties, whereas the second involves contour extraction in a single image.
RIS
TY - CPAPER TI - Variational Inference for Visual Tracking AU - Jaco Vermaak AU - Neil D. Lawrence AU - Patrick Pérez BT - PY - DA - ED - ID - pmlr-v-vermaak-variational03 PB - PMLR SP - 773 DP - PMLR EP - 780 L1 - UR - http://inverseprobability.com/publications/vermaak-variational03.html AB - The likelihood models used in probabilistic visual tracking applications are often complex non-linear and/or non-Gaussian functions, leading to analytically intractable inference. Solutions then require numerical approximation techniques, of which the particle filter is a popular choice. Particle filters, however, degrade in performance as the dimensionality of the state space increases and the support of the likelihood decreases. As an alternative to particle filters this paper introduces a variational approximation to the tracking recursion. The variational inference is intractable in itself, and is combined with an efficient importance sampling procedure to obtain the required estimates. The algorithm is shown to compare favourably with particle filtering techniques on a synthetic example and two real tracking problems. The first involves the tracking of a designated object in a video sequence based on its colour properties, whereas the second involves contour extraction in a single image. ER -
APA
Vermaak, J., Lawrence, N.D. & Pérez, P.. (). Variational Inference for Visual Tracking. , in PMLR :773-780

Related Material