Fetal Ultrasound Image Segmentation

Abstract

The thesis explains development of a time-efficient method for segmentation in fetal ultrasound images so as to recognize femur contour in fetal ultrasound images and refine its shape. The main objective of the project was to develop segmentation of womb and fetus using fetal ultrasound images. Ultrasound segmentation produces detailed visualization of the underlying anatomy of the fetus in real time which can also recognize dynamically varying structures of the fetal body and the motion of the fetus in the mother’s uterus. These are the factors which could be got by ultrasonograph segmentation, on which fetus can be distinguished as normal or abnormal. This thesis explains the development of enhancement, thresholding and segmentation algorithms as the ultrasound diagnosis highly depends on the quality of the data. The developed algorithms segment the fetus which helps in analyzing the anatomical structure of the fetus.
Keywords: Grayscale images, adaptive thresholding, median filtering, histogram equalization, morphological operation, Prewitt’s edge detection