Color textured MRI segmentation based on Co-occurrence features using principal component analysis (PCA)‎

Abstract EN

Image segmentation is a primary step in many computer vision tasks.

Although many segmentation methods based on either texture or color have been proposed in the last decades, there have been only few approaches combining both of these features.

In this paper the Principal Component Analysis (PCA) of the counted co-occurrence featured images has been introduced as a new segmentation algorithm for differentiating tumor tissue.

Coloring process is also performed by selecting three PCA’s bands.

An encouraging results obtained by combining different PCA’s bands.

This method reveals the importance of applying context and textural properties for the segmentation process.

The results showed that utilizing different multivariate statistical techniques (e.g.

gray level Co-occurrence matrices and Principal Component Analysis technique); based on color-band, mixing method has can be used to isolate the tumor tissues from other brain tissues.