Texture-Based Image Enhancement using Gabor Filters and Morphological Operations

As a field image enhancement has evolved due to an inherent human want for better image quality. Traditional methods apply linear scaling functions operating on image attributes such as contrast and brightness whereas it is often necessary to amplify specific localized elements as edges in images. This need has resulted in a search for better methods in efforts to boost these local characters in as much as the general image is concerned. Thus, this paper aims to propose an edge- and texture-aware image enhancement framework based on full-depth Gabor filters and morphological operation that can overcome the above drawbacks and gain the desired image texture complexities and brightness. Hence the method proposed in this work aims at preserving edge and texture at the same time to get an image that looks natural and has improved aesthetics. This bank of filters is mostly useful to obtain small details of texture and sharpness. An inverse gamma transform is then applied to the image to reduce gamma distortions in the image, while another process known as depth-of-defocus is to determine the edges of the texture image. These are the detected edges used in the coarse refinement stage. Morphological operations are employed to populate and repaint the elaborate structures of this image to fit up. An experimental analysis of the proposed method is carried out by conducting experiments on several structures of images including images with different textured content, different levels of brightness, and noise. This result is generalized with the help of quantitative measures and qualitative analysis on PSNR, SSIM, FSIM, and GMSD parameters.