This thesis has two related goals: the first involves the concept of self-similarityof images. Image self-similarity is important because it forms the basis for manyimaging techniques such as non-local means denoising and fractal image coding.Research so far has been focused largely on self-similarity in the pixel domain.That is, examining how well different regions in an image mimic each other. Also,most works so far concerning self-similarity have utilized only the mean squarederror (MSE).In this thesis, self-similarity is examined in terms of the pixel and wavelet representationsof images. In each of these domains, two ways of measuring similarityare considered: the MSE and a relatively new measurement of image fidelity calledthe Structural Similarity (SSIM) Index. We show that the MSE and SSIM Indexgive very different answers to the question of how self-similar images really are.The second goal of this thesis involves non-local image processing. First, ageneralization of the well known non-local means denoising algorithm is proposedand examined. The groundwork for this generalization is set by the aforementionedresults on image self-similarity with respect to the MSE. This new method is thenextended to the wavelet representation of images. Experimental results are givento illustrate the applications of these new ideas.
【 预 览 】
附件列表
Files
Size
Format
View
Self-Similarity of Images and Non-local Image Processing