Automated detection of fetal cardiac structure from first-trimester ultrasound sequences

Abstract

The structure detection of the first-trimester fetal heart is important for the diagnosis of fetuses, which is difficult due to the small size of the first-trimester fetal heart and the low signal-to-noise ratio of the ultrasound imaging. In the paper, a novel method is proposed for the automated detection of the fetal cardiac structure from first-trimester ultrasound images. First the region of interest for the fetal heart is automatically selected based on the established motion summation image. Then a despeckle method is proposed to simultaneously suppress the speckle noise according to its statistical property and emphasize the motion information which is important for the next detection. Finally an active appearance model is designed for the fetal heart and accomplishes the structure detection. Experiments on 258 ultrasound images verify the effectivity of our proposed method.