Fraction B-spline Wavelet And Image Processing

In this paper, we first introduce the fractional B-spline wavelets proposed by Blu and Unser, and discuss their properties and construction method. Then we extend the one-dimension fractional B-spline wavelet transform to two-dimension, and implement the image decomposition and reconstruction, based on the fractional spline wavelet transform. We analyze the statistics characteristics and distributions of fractional B-spline wavelet coefficients of images, for different values of a . The image denoising techniques based on fractional B-spline wavelet are studied and compared with classical wavelet denosing techniques from MSE and PSNR. Then, we give a new image denoising method based residual error correction model, which improves the quality of image denoising. Finally, we propose the concept of image energy eigenvector, and implement image classification.