Automated denoised ultrasound carotid artery image segmentation using curvelet threshold decomposition

In this paper we propose denoising Common Carotid Artery (CCA) B-mode ultrasound images by a decomposition method to curvelet thresholding and segmentation inevitably the intima-media thickness and adventitia boundary. By decomposition, the local geometry of the image, its direction of gradients is conserved. The components are joint into a distinct vector valued function, thus eliminates noise patches. The dual threshold is smeared to remove speckle noise in the image. The denoised image is segmented by active contour without specifying seed points. Joined with level set theory, they offer sub-regions with continuous boundaries. The deformable contours match to the shapes and motion of objects in the images. A curve or a surface under limits is established from the image with the goal that it is drawn in to the required features of the image. Region based and boundary based information are integrated to achieve the contour. The method treats the multiplicative speckle noise in objective and subjective quality measurements and thus leads to improved segmented results. The proposed denoising method gives better performance metrics compared with other state of art denoising algorithms.