Filtering and Left Ventricle Segmentation of the Fetal Heart in Ultrasound Images

摘要

In this paper, we propose to use filtering methods and a segmentation algorithm to analyze the fetal heart in ultrasound images. Since speckle noise makes difficult the analysis of ultrasound images, the filtering process becomes a useful task in this application. The filtering techniques considered in this work assume that the speckle noise is a random variable with a Rayleigh distribution. We use two multiresolution methods: one based on wavelet decomposition and the other based on the Hermite transform. The filtering process is used as a way to strengthen the performance of the segmentation task. For the wavelet-based approach, a Bayesian estimator at subband level for pixel classification is employed. The Hermite method computes a mask to find those pixels that are corrupted by speckle. We picked out a method based on a deformable model or "snake" to evaluate the influence of the filtering techniques in the segmentation task. We selected the left ventricle in fetal echocardiographic images as structure of analysis. Quantitative evaluation is addressed to assess the performance of the filtering process and the segmentation task.