Conference Paper

A stochastic approach to ultrasound despeckling

Dept. of Electr. & Comput. Eng., Virginia Univ., Charlottesville, VA
DOI: 10.1109/ISBI.2006.1624892 Conference: Biomedical Imaging: Nano to Macro, 2006. 3rd IEEE International Symposium on
Source: IEEE Xplore

ABSTRACT A novel stochastically driven filtering method to despeckle B mode ultrasound images is presented. This method is motivated by viewing the pixel values as a stochastic process and removing outliers, where outliers are defined by local extrema. These outliers are removed by local averaging. This produces another image with new outliers (local extrema) and the process is iteratively repeated. With each iteration homogeneous regions become smoother while edges that defined these regions remain preserved. To evaluate the performance of our proposed method in satisfying these two opposing goals we develop a modified Fisher discriminant contrast metric. Larger values of this metric indicate better performance in reducing each intraregion or intraclass variance and increasing the difference of interregion or interclass means

  • [Show abstract] [Hide abstract]
    ABSTRACT: A new speckle reduction method, i.e., an adaptive nonlocal means based on fuzzy domain (FANLM), is proposed for polarization imaging despeckling. With this method, speckle is removed by an adaptive nonlocal means in Laplacian pyramid domain. The proposed FANLM method calculates the similarity measure in the fuzzy domain, which can make the similarity measure more precise and accurate. In view of the similarity neighborhood selection problem, an adaptive selection algorithm is proposed. The performance of the proposed FANLM method has been compared with the classic methods. Experimental results show that the visual quality and evaluation indexes outperform the other methods. The researches have important reference values in removing speckle noise.
    Optik - International Journal for Light and Electron Optics 04/2014; 125(7):1833–1836. DOI:10.1016/j.ijleo.2013.09.037 · 0.77 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Ultrasound imaging is one of the most important instruments of modern medical imaging modalities. However, the usefulness of ultrasound imaging is degraded due to the presence of a signal-dependent noise known as speckle. Recently, lots of algorithms have been proposed for despeckling. Unfortunately, few attentions have been paid on time series of ultrasound images like echocardiography. In this paper, we address this problem by developing a time series non-local means (NLM) filter algorithm. A distance measure relevant to a speckle model is firstly introduced to take place of the Gaussian-weighted Euclidian distance according to a Bayesian formulation. By taking the information along the temporal axis into account, we further extend the NLM filter from single frame to image time series. To lighten the computational burden, a blockwise approach and a pre-classification process are used to accelerate the algorithm. In order to evaluate our method, experiments are conducted on both synthetic and in vivo ultrasound images. Experiments show that the proposed method achieves satisfactory results in terms of removing the speckle and preserving the edges and image details, compared with the state-of-the-art methods.
    Neurocomputing 01/2014; 124:120-130. DOI:10.1016/j.neucom.2013.07.022 · 2.01 Impact Factor
  • Source
    Ultrasound Imaging, Edited by João Miguel Sanches, Andrew F. Laine, Jasjit S. Suri, 01/2012: pages 283-297; Springer., ISBN: 978-1-4614-1179-6