Fig. 8
From: Two-step verification of brain tumor segmentation using watershed-matching algorithm
Experimental result of SIFT segmentation (from top to bottom). Finding features of input MR brain tumor image, finding keypoints; the density of green signs and the blue signs on the tumor area indicate the tumor region; keypoints are localized, and then, find matching and lastly find 89 matches between the input and the output image, thus confirming the tumor area from the binary image