Fractional-order High-boost Filtering for Textural Improvement of Images using Relative Spatial Entropy Quartiles

Abstract

Textural segmentation and its usage for region-wise image quality improvement unfold a new chapter for texture-dependent image processing in association with fractional order calculus (FOC). Along with intensity variation, texture variation is also equally important for human as well as machine vision to discriminate between surfaces and objects even having the same intensity. Most of the vision applications deal with intensity-wise segmented frames as their raw input. The power of textural analysis along with conventional intensity-based processing can enhance the system's capability in a remarkable manner. To address the textural nature of the image and for imparting texture-dependent image restoration or enhancement fractional-order high-boost filtering (FoHBF) the framework is essentially relevant irrespective of the image domain. Spatial entropy quantile-based textural segmentation and region-wise FoHBF is employed in this paper for imparting total quality enhancement, especially for remotely sensed images.