Efficient Segmentation of the Foetal Ultrasound Image Using Smoothing Algorithm

Ultrasound (US) imaging is the modality of choice in many clinical applications compared to other imaging modalities, such as computed tomography (CT). US images are patient-specific, operator-dependent, and machine specific. However these US images are affected by signal dropouts, artefacts, missing boundaries, attenuation, shadows, and speckle, making US one of the most challenging modalities to work with. Earlier work presents qualitative and quantitative segmentation evaluation of the representative selection method submitted to Challenge US: Biometric Measurements from Fetal Ultrasound Images. In this a total of five teams submitted their results to the fetal head sub-challenge and two teams to the fetal femur sub-challenge, including one team who attempted both. The results of the fetal head sub-challenge shows good performance whereas a fetal femur sub-challenge faces the problem in solving a very hard segmentation problem, since the object of interest has strong appearance changes within the object. To deal with this problem the present work proposes smoothing algorithm to smoothen the other elongated objects are present around the femur bone. This algorithm works by computing the optimal fixed range bandwidth in the US image. Finally the proposed work is experimentally evaluated and thus obtains improved result when compare with the existing system.