An Image Performance Against Normal, Grayscale and Color Spaced Images

Abstract

AbstractGenerally, an image matching belongs to comparing the two images, with the simple concept i.e. When the two images match or comparable and how can this similarity be measured? Fast and robust feature detection and image matching have always been a very major and challenging task in itself along with the applications. In this paper, we are using normal, grayscale and LAB color spaced images and measure the recital of contrasting approaches for image corresponding, i.e., SIFT, SURF, and ORB. For this purpose, we manually transform original images into grayscale and LAB color spaced images and compute all the parameters on the basis of which evaluation is done such as the total of distinct points in images, the match-up percentage. By this, we will show that which algorithm works best and more robust against each kind of image.KeywordsImage matchingScale-invariant feature transform (SIFT)Speed Up Robust Feature (SURF)Oriented FASTRotated BRIEF (ORB)