Selection of the Best Despeckle Filter of Ultrasound Images

Ultrasound imaging is considered as the largest medical imaging modalities even it suffers from despeckle noise. While there are dissimilar despeckling techniques to remove noise, they are not efficient with all images. In addition, the physician will not be able to select the best technique manually. The four despeckling techniques are; linear filter, non-linear filter, diffusion filter and wavelet filter. This paper implements these techniques on a specific dataset. The results are evaluated based on the expertise opinion. Moreover, a comparison is conducted between the expertise opinion and the extracted features from both original and despeckles images. We apply parallel coordinate to visualize the extracted features before and after applying best despeckle techniques to know the dominant features that lead to choose the suitable technique. The results show that there are dominant features like contrast, correlation, entropy, mean and variance