Robust optic-flow estimation with bayesian inference of model and hyper-parameters

Authors:
P. Héas;C. Herzet;E. Mémin
Affiliations:
Fluminance, INRIA Rennes-Bretagne Atlantique, Rennes, France;Fluminance, INRIA Rennes-Bretagne Atlantique, Rennes, France;Fluminance, INRIA Rennes-Bretagne Atlantique, Rennes, France
Venue:
SSVM'11 Proceedings of the Third international conference on Scale Space and Variational Methods in Computer Vision
Year:
2011

Citing 12
Cited 0

Constrained Restoration and the Recovery of Discontinuities

IEEE Transactions on Pattern Analysis and Machine Intelligence
Bayesian interpolation

Neural Computation
The robust estimation of multiple motions: parametric and piecewise-smooth flow fields

Computer Vision and Image Understanding
An Introduction to Variational Methods for Graphical Models

Machine Learning
Consistent Segmentation for Optical Flow Estimation

ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
A Multigrid Platform for Real-Time Motion Computation with Discontinuity-Preserving Variational Methods

International Journal of Computer Vision
Discrete Orthogonal Decomposition and Variational Fluid Flow Estimation

Journal of Mathematical Imaging and Vision
Over-Parameterized Variational Optical Flow

International Journal of Computer Vision
Learning Optical Flow

ECCV '08 Proceedings of the 10th European Conference on Computer Vision: Part III
An iterative image registration technique with an application to stereo vision

IJCAI'81 Proceedings of the 7th international joint conference on Artificial intelligence - Volume 2
Bayesian model selection for optical flow estimation

Proceedings of the 29th DAGM conference on Pattern recognition
Bayesian and regularization methods for hyperparameter estimation in image restoration

IEEE Transactions on Image Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Selecting optimal models and hyper-parameters is crucial for accurate optic-flow estimation. This paper solves the problem in a generic variational Bayesian framework. The method is based on a conditional model linking the image intensity function, the velocity field and the hyper-parameters characterizing the motion model. Inference is performed at three levels by considering maximum a posteriori problem of marginalized probabilities. We assessed the performance of the proposed method on image sequences of fluid flows and of the "Middlebury" database. Experiments prove that applying the proposed inference strategy on very simple models yields better results than manually tuning smoothing parameters or discontinuity preserving cost functions of classical state-of-the-art methods.