Paper Details: Downloads: 667
Serial Number: P1150724001
Title: Medical Image Registration using Regions Matching with Invariant Geometrical Moments
Authors: G. Khaissidi and M. Karoud and H. Tairi and A. Aarab
Abstract: Image registration is an important problem for image processing and computer vision with many proposed applications in medical image analysis. Image registration techniques attempt to map corresponding features between two images. In this paper we propose an approach for medical images registration characterized by their great heterogeneity and complexity. It is carried out through the extraction of interest regions or points of the both images source and destination. The correlation technique ZNCC (Zero mean Normalized Cross Correlation) combined with relaxation is used for matching the interest points. In region based registration approach, the image segmentation involves growing regions from individual pixels using a stack-based approach. The matching between the extracted regions is established with a new descriptor that involves three parameters: color, position, and shape. The shape is characterized by the invariant geometrical moment. The least means square is used to estimate the registration parameters. The whole results obtained are rather satisfactory and very encouraging.
Keywords: registration, medical imagery, interest points, interest regions, matching.
Journal/Conference: International Journal of Graphics, Vision and Image Processing
Volume: 8
Issue: 2
Submission Date: 7/1/2007 12:00:00 AM
Review Date: 6/1/2008 12:00:00 AM
Publishing Date: 7/1/2008 12:00:00 AM
Article Downloads: 667
Download:

Facebook