Research Of Image Enhancing Of Vein Based On Bandelet Transform

With the development of society and the progress of science and technology, biological recognition technology (such as the iris, hand vein, fingerprints, etc.) becomes the hotspot of research. Among them, the vein recognition has been giving more and more attention for the unique advantages, which is difficult to counterfeit and non-touch acquisition. But, for the influences of the equipment and lighting factors, the contrast of vein image is lower, and the vein images has more noise which bring a lot of inconvenience to the recognition. So the vein image de-noising and enhancement are very necessary before the recognition of vein image.Owing to the low-entropy and multi-resolution of wavelet,wavelet is widely used in image denoising and compression.But in high dimensional, wavelet fails to represent the geometric features such as multi-directions edges and textures in vein image effectively. Bandelet could overcome the insufficiency of wavelet in processing high dimensional signals, and achieve the optimal approximation rate for geometrically regular images。 This paper take the image denoising and enhancement algorithm based on Bandelet thesis which could take advantage of the Bandelet in representing the image geometrical features and preserving the edges of image.Notice that wavelet suffers from some shortcomings,shift variance and lack of directionality,results in Bandelets which are based on wavelet domain affected by the wavelet shortcomings.In view of the above,this paper takes the denoising Algorithm based on DTCWT-Bandelet transform which generate the orthogonal Bandelets based on DTCWT domain.A comparison is made with those results obtiand with Wavelets, Bandelet, DTCWT, and DTCWT-Bandelets for denoising process with addictive white Gaussian noise image by Garrote thresholding function,which shows that DTCWT-Bandelets could outperform the wavelets and Bndelet in terms of subjective and objective measures.Moreover,A image enhancement arithmetic based on the DTCWT-Bandelet transform with non linea enhancement function which could suppress the noise and enhance the image at the same time is presented for the noise image in this paper. A comparison of Wavelets, Bandelet, DTCWT, and DTCWT-Bandelets for denoising was made.and results of experiments show that the methods based on DTCWT-Bandelet offers significant improved performance in emphasizing features and reducing distortion of edges.