Fetal ultrasound image segmentation using fuzzy clustering methodology

Identifying object boundaries in ultrasound images is a difficult task for even the most experienced physicians. This is due to the inherent noise present in ultrasound images introduced during creation of these images. Segmenting ultrasound images is very much different from segmenting normal images due to the noise. This thesis is aimed at developing a segmentation algorithm involving a fuzzy clustering methodology to segment fetal ultrasound images. This algorithm uses a fuzzy clustering methodology to segment any given fetal ultrasound image into a user specified number of segments. The process starts with a default number of clusters and follows an unsupervised learning process to output a segmented image with the specified number of gray levels. The resulting images are segmented and superimposed with edges, facilitating improved readability and analysis of fetal ultrasound images to determine the growth of the fetus without losing important detail in the image.