Development of medical image segmentation algorithms for focal liver lesion extraction in MR images

• Κύριος συγγραφέας: Λύγκρης, Γεώργιος
• Άλλοι συγγραφείς: Καγκάδης, Γεώργιος
• Μορφή: Thesis
• Γλώσσα: English
• Έκδοση: 2020
• Θέματα: Segmentation; Algorithms; Liver lesions; Τμηματοποίηση; Αλγόριθμοι; Ήπαρ; Ηπατικές αλλοιώσεις

In this postgraduate thesis, the technique of segmentation was studied that plays a dominant role in the field of medical image analysis and processing. The medical images that were processed came from an MRI scan and depict liver lesions (benign and malignant). The anatomy and physiology of the liver in both a healthy and pathological environment are described in detail. The segmentation algorithms of K-Means, Fuzzy C-Means, Markov Random Fields and were studied and the experimental results are presented with the conclusion and concerns. The algorithms were first tested on phantom images and then applied to medical images by Μagnetic Resonance Imaging. Finally, the Jaccard Index, which is a criterion for comparing algorithms, was calculated. The algorithms were developed, modified and executed in MATLAB programming environment.