Medical Image Segmentation Using a Genetic Algorithm
[report]
Payel Ghosh
2000
unpublished
Advances in medical imaging technology have led to the acquisition of large number of images in different modalities. On some of these images the boundaries of key organs need to be accurately identified for treatment planning and diagnosis. This is typically performed manually by a physician who uses prior knowledge of organ shapes and locations to demarcate the boundaries of organs. Such manual segmentation is subjective, time consuming and prone to inconsistency. Automating this task has
more »
... found to be very challenging due to poor tissue contrast and ill-defined organ/tissue boundaries. This dissertation presents a genetic algorithm for combining representations of learned information such as known shapes, regional properties and relative location of objects into a single framework in order to perform automated segmentation. The algorithm has been tested on two different datasets: for segmenting hands on thermographic images and for prostate segmentation on pelvic computed tomography (CT) and magnetic resonance (MR) images. In this dissertation we report the results of segmentation in two dimensions (2D) for thermographic images; and two as well as three dimensions (3D) for pelvic images. We show that combining multiple features for segmentation improves segmentation accuracy as compared with segmentation using single features such as texture or shape alone. ii To my father who instilled the love of mathematics in me. iii Acknowledgments I would first like to thank my advisor, Dr. Melanie Mitchell, not only for her constant guidance and advice, but also her consideration for my personal situation which helped me complete part of this research work off-campus. I would like to express my sincere thanks to Dr. James Tanyi from OHSU. It was due to his immense help that the data from OHSU was acquired in a matter of weeks despite administrative issues. I am thankful to him and Dr. Arthur Hung for providing manual segmentations for the CT and MR data sets. I am also grateful to Dr. Judith Gold from Temple University for providing the second dataset of thermographic images. The dataset proved to be ideal for testing and evaluating the developed algorithm. I would also like to thank all my committee members for their valuable comments and feedback on my work.
doi:10.15760/etd.25
fatcat:hkguenpkije2xgqd3256ic7caa