Two-Fold Domain Image Denoising

Abstract

Image denoising has become an important sub domain of image processing with its applications and uses far reaching and eternally evolving. It is an important and rather an unescapable pre-processing step used in various applications such as the ones used to compress an image, segmentation fusion, and recognizing of objects. Here, an image denoising technique is defined in a bi-fold model. The denoising process is performed in spatial domain and frequency domain. The main objective is to come up with a technique that is not only efficient but is also cost and computation effective. Existing techniques have also been working on dual domain but leave something to be desired when it comes to computation time. The proposed technique of Two-Fold Domain Image Denoising makes use of Guided Image Filtering technique in the spatial domain where it preserves the edges and unwanted data is eliminated, orthonormal wavelet thresholding is applied in wavelet domain where the contrast is preserved. The whole process runs in three iterations which greatly reduce the computational time. A detailed comparison between existing techniques and proposed technique is also listed in the document to illustrate the efficiency of the proposed model.