Decomposition Methods: A Comparative Analysis using Entropy Feature from Fundus Images

Abstract

Glaucoma is an incurable eye disease; it destroyed the optic nerve head due to continuing increase in the fluid pressure in the eye. In this paper different image decomposition methods are compared and examined. The proposed work uses mainly three decomposition methods, namely, bi-dimensional empirical mode decomposition (BDEMD), two dimensional empirical wavelet transform (2EWT) and two dimensional variational mode decomposition (2DVMD). Glaucoma and normal images are decomposed by these methods and entropy features are extracted from the decomposed sub band images. The percentage variation in entropy features (PVIEF) are calculated from the extracted entropy features using decomposition method for normal and glaucoma images. The calculated PVIEF are used to compare the three decomposition methods for normal and glaucoma images. The obtained results put forward that the PVIEF extracted from 2DVMD are highest among all the three decomposition methods. Hence, 2DVMD has the highest ability for detection and classification of glaucoma and outperforms over 2DEWT and BDEMD.