A Markov-Random-Field Based Filter for Speckle Reduction in Ultrasound Imagery

摘要

The coherent nature of ultrasound imaging inherently produces the notorious signal-dependent speckle noise. Recently, a novel approach based upon embedding the statistical and physical properties of speckle patterns into a Markov-random-field (MRF) framework was developed and demonstrated by the authors in the context of synthetic-aperture radar imaging. The contributions of this work are twofold. First, the MRF approach is extended to include a pseudo maximum-likelihood estimator of a key model parameter, making the approach fully autonomous. Second, the capability of the extended approach, called the modified MRF-based conditional-expectation approach (MRFCEA), in denoising real ultrasound imagery is demonstrated. The proposed MRFCEA approach offers superior performance over existing methods by reducing speckle noise without compromising the spatial resolution. In addition, MRFCEA is autonomous, contrary to existing methods such as the enhanced-Frost or the modified-Lee, which require user's input.