DMCA
Dense Image Registration through MRFs and Efficient Linear Programming (2008)
| Citations: | 75 - 37 self |
BibTeX
@MISC{Glocker08denseimage,
author = {Ben Glocker and Nikos Komodakis and Georgios Tziritas and Nassir Navab and Nikos Paragios},
title = { Dense Image Registration through MRFs and Efficient Linear Programming},
year = {2008}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
In this paper we introduce a novel and efficient approach to dense image registration, which does not require a derivative of the employed cost function. In such a context the registration problem is formulated using a discrete Markov Random Field objective function. First, towards dimensionality reduction on the variables we assume that the dense deformation field can be expressed using a small number of control points (registration grid) and an interpolation strategy. Then, the registration cost is expressed using a discrete sum over image costs (using an arbitrary similarity measure) projected on the control points, and a smoothness term that penalizes local deviations on the deformation field according to a neighborhood system on the grid. Towards a discrete approach the search space is quantized resulting in a fully discrete model. In order to account for large deformations and produce results on a high resolution level a multi-scale incremental approach is considered where the optimal solution is iteratively updated. This is done through successive morphings of the source towards the target image. Efficient linear programming using the primal dual principles is considered to recover the lowest potential of the cost function. Very promising results using synthetic data with known deformations and real data demonstrate the potentials of our approach.
Keyphrases
dense image registration efficient linear control point optimal solution known deformation real data registration cost large deformation image cost search space towards dimensionality reduction local deviation discrete model primal dual principle arbitrary similarity measure target image neighborhood system deformation field employed cost function synthetic data successive morphings high resolution level dense deformation field multi-scale incremental approach small number interpolation strategy registration problem registration grid image registration efficient approach discrete sum efficient linear programming discrete approach smoothness term cost function promising result
