Publications

Abstract : In this chapter, a texture-dependent optimal fractional-order adaptive filtering is proposed for quality improvement of the remotely sensed dark satellite images. Each image is usually composed of diverse variations in texture. To identify and address the texture-based variation in any image, it is highly desired to identify the different kinds of texture constituents present in the image. This objective can be easily attained by texture-based segmentation. Texture based segmentation is usually performed by using the spatial information content present in the spatial texture map of the image under consideration. Spatial entropy based texture map is computed in this work. Later, the grouping the varying textural behavior in multiple classes is done. It makes easy to process the different textural regions separately. In this manner, various classes of relative (or normalized) texture are tried to process individually. For this purpose, optimal fractional-orders are required to be computed for each class of relative texture present in the image. A dedicated optimization-based fractional-order filtering framework has been drafted for fulfilment of the objective. To impart a high-level meta-heuristic intelligence for this optimal framework, in this chapter, the collective excellence of two diversely-behavioral approaches is obeyed in a collective mode. In this context, a hybrid intelligence of Grey-Wolf Optimizer (GWO) achieved by using Cuckoo Search Algorithm (CSA) for applying optimal fractional. As a whole a fusion framework is presented by associating all individual interim channels. A rigorous comparative experimentation is performed and visual as well as numerical analyses are presented in this chapter. The excellence of the proposed approach is underlined when compared w.r.t. the state of the art image enhancement approaches.