| In recent years, with the development of ultrasound imaging, it has become one of the important diagnostic methods in medical imageology. However, the imaging equipments,imaging principles and other reasons will lead to the speck noise appears in its final spot imaging.Medical ultrasound images produced multiplicative speckle noise which is signal-dependent in the process of imaging. Speckle noise significantly degrades the distinguished nature of ultrasound images and seriously affects the quality of image especially causes the loss of the detail information and vagues the edge of images. All these are increased the difficulties to separate the noise from the signal which has important theoretical significance and application value.The multiplicative speckle noise can be converted to an addive noise by log-transforming the image. But in the process of research the log-transforming of ultrasound images, there are great significances of how to suppress speckle multiplicative noise and to retain and enhance the edge features and detail information of the image.This paper is based on speckle reducing anisotropic diffusion to establish a noise-removal model. Considering the concentration of wavelet coefficients, we introduced the wavelet transform to divide homogeneous region and proposed a division method to classify the wavelet coefficients. By proposing a high robust diffusion coefficient, adding diagonal gradient information when calculate gradient operator and Laplacian operator and adding fidelity term in the algorithm to improve anisotropic diffusion speckle noise and form an efficient wavelet-based anisotropic diffusion medical ultrasound image noise-removal algorithm. Experimental results show that the method on ultrasound images of the suppression poposed in this paple is feasible. Comparing with other types of traditional and classical filter methods, this method obtain better results of quantitative analysis and better visual effects. |
PDF Full Text Request