Automatic segmentation of brain MRI images and tumor detection using morphological techniques

Automatic segmentation in image processing is the method of isolating an image into mutually exclusive regions. During the processing of brain MRI images, segmentation is considered as the essential and crucial step because of the diverse image content, artifacts and disordered objects, non-uniform object texture and other issues. In this paper, automatic segmentation by morphological operations is implemented and the result is compared with other segmentation techniques like Expectation maximization and Fuzzy C-Means with reference to performance measures and processing time. The performance measures such as Jaccard Distance, Dice Coefficient, False Positive Ratio, and False Negative Ratio are used for comparison. The experimental results clarify the effectiveness of segmentation algorithms in terms of quality and accuracy in minimal execution time. The morphological segmentation is found to be fast and effective in automatic segmentation of brain MR images.