Automated image registration by maximization of a region similarity metric
✍ Scribed by Zhi-Pei Liang; Hao Pan; Richard L. Magin; Narendra Ahuja; Thomas S. Huang
- Publisher
- John Wiley and Sons
- Year
- 1997
- Tongue
- English
- Weight
- 204 KB
- Volume
- 8
- Category
- Article
- ISSN
- 0899-9457
No coin nor oath required. For personal study only.
✦ Synopsis
This article presents a robust algorithm for automated problem, new methods for automated registration of multimodalregistration of images related by rigid-body transformations. This ality images continue to appear [9,10]. This article describes angorithm uses a new region-based similarity metric, which enables other such method. A key contribution here is a new region accurate registration of images of large contrast differences. Region similarity metric which can effectively handle images with large segmentation required by the metric is accomplished using a contrast differences. Region segmentation required by the metric multiscale segmentation algorithm, and minimization of this metric is is accomplished using a multiscale segmentation algorithm. Minidone using the Powell direction set method. Experimental results are mization of the metric is done using the Powell direction set presented to demonstrate that the algorithm is effective for aligning method. Details of this technique are provided in the subsequent images from single or multiple imaging modalities without the use of section. Sample experimental results are shown in Section III to any fiducial markers.