A Novel Approach for Extricating Information from Speckle Noise Corrupted Color Brain MRI Images Using Morphological Enhancement Followed by Watershed and Region Growing Based Segmentation

Segmentation of an image is performed to extricate desired information for analysis. Brain MRI images require segmentation for extricating useful information for detecting tumors and disease related to white matter like multiple sclerosis for diagnosis of disease. The segmentation methodology may be adversely affected when performed on images affected with noise. This paper offers a novel approach for segmentation of speckle noise corrupted brain MRI images using noise filtering, enhancement using morphology, watershed transform followed by region growing technique for information extraction. The proposed method provides finer results as compared to mean shift and K means color image segmentation techniques as well as fuzzy c means clustering with improved robustness.