Publications

Abstract : In this contemporary era, technological dependencies on digital images are indispensable. Quality enhancement is an obligatory part of image pre-processing, so that desired information can be harvested efficiently. The varying texture in any image contributes to the information about the structural arrangement of surface content of the captured scene. Fractional-order calculus (FOC) and its related optimally ordered adaptive filtering are quite appreciable. Especially for texture preserved image quality enhancement, analytical strength of FOC is latently too valuable to be casually dismissed. No any closed-form free-lunch theory survives for evaluating the required fractional-order for overall quality enhancement because non-linear features of images from diverse domains require highly adaptive on-demand processing. Hence, texture preserved image quality enhancement can be considered as an NP-hard problem, where there isn’t an exact solution that runs in polynomial time. Thus, by the virtue of evolutionary algorithms along with their associated swarm intelligence, a near-exact solution can be attained. Memetic hybridization of cuckoo search optimizer (CSO) and sine-cosine optimizer (SCO) for this purpose is the core contribution in this chapter. In this chapter, to support the theoretical discussion in the context of the fundamentals behind CSO and SCO, their mathematical beauty of convergence is also highlighted which itself has resulted from the balance exploration and exploitation behavior. A novel texture-dependent objective function is also proposed in this chapter for imparting the patch-wise overall texture preserved image quality enhancement. Finally, the comparative analysis of results illustrates the superior capability of the proposed approach.