No abstract

No abstract

Ultrasound imaging has been considered as the most powerful techniques for imaging organs and soft tissue structures in the human body. However its main limitation is its poor quality of images which are degraded by speckle noise. Speckle is a multiplicative form of noise which is inherent in ultrasound imaging but carries some useful information which should be filtered out without losing the features in an image. Fuzzy logic has the capability of handling the input that is approximate rather than fixed and exact. Since the ultrasound images have fuzziness in nature caused by speckle noise, vague edges and boundaries, fuzzy logic can be considered as the simple way to arrive at a definite output based on vague, ambiguous, imprecise or noisy input information. In this paper, a novel fuzzy applied filter has been designed for speckle noise reduction based on directional differences in which noise affected pixels are classified depending on the magnitude of noise. Appropriate filters are applied on the corrupted pixels which gives appreciable improvement both visually and in PSNR values.

Speckle noise reduction is an important preprocessing stage for ultrasound medical image processing. In this paper, a despeckling algorithm is proposed based on non-subsampled contourlet transform. This transform has the property of high directionality, anisotropy and translation invariance, which can be controlled by non-subsampled filter banks. This study aims to denoise the speckle noise in ultrasound images using adaptive binary morphological operations, in order to preserve edges, contours and textures. In morphological operations, structural element plays an important role for image enhancement. In this work, different shapes of structural element have been analysed and filtering parameters have been changed adaptively depending on the nature of the image and the amount of noise in the image. Experimental results of proposed method for natural images, Field II simulated images and real ultrasound images, show that the proposed method is able to preserve edges and image structural details compared with existing methods.

No abstract