Image Contrast Enhancement by Homomorphic Filtering based Parametric Fuzzy Transform

摘要

In this paper, a new image contrast enhancement technique by taking the advantages of Homomorphic decomposition and fuzzy transform is been employed. For increasing the clarity of the image generally, histogram or Retinex theory based algorithms were used. These procedures worked on enhancing the reflectance layer by ignoring illumination, which is not a better strategy and leads to poor results. Fuzzy based image enhancement approach makes use of illumination by omitting reflectance. In our proposed algorithm, Homomorphic decomposition is used for getting the exact illumination image from the value layer of HSV (Hue, Saturation, Value) image. Next, the parametric fuzzy transform is employed to enhance the image by updating its membership functions and thereby smoothing the luminance layer. Finally, a weighted image is generated by pixel neighborhood property for preserving the image details. The results show the profoundness of the algorithm in terms of its clarity and complexity even for nonuniform illumination images.