An efficient two step algorithm for despeckling the ultrasound image

Abstract

The speckle noise presence in ultrasound images is a critical concern in medical image processing. It degrades the important features captured in an image and decreases the physician’s capacity to understand the image accurately. In recent years, numerous techniques have been proposed to de-noise the ultrasound images. In this paper, a new speckle noise removal algorithm has been proposed for medical ultrasound images. Based on the concepts of fuzzy logic and Coefficient of variation, the proposed algorithm first classifies the image area into three different regions such as homogeneous, edge and detail region. Next, average filter, median filter and an adaptive mean filter are employed to partition the unwanted noise from the pixels of different regions. Filter selection depends on the features of a region. The proposed algorithm develops image quality by removing maximum unwanted noise while protecting the important image details.