Nonlinear cost optimization scheme for feature segmentation in second trimester fetal images

The shape sensitive derivative class separable segmentation scheme for the ultrasound fetal images is presented in this paper. The energy cost function is optimized with topological asymptotic expansion for feature extraction. The speckle present in the image has been removed by improved iterative median filter. The algorithm has been used for segmenting nasal bone and nuchal translucency in second trimester fetal images and validated experimentally by evaluating dice similarity coefficient, tanimoto index and overlap index.