Author, Institution: Jewel Sengupta, Kaunas University of Technology
Science area, field of science: Technological Sciences, Informatics Engineering, T007
Research supervisor: Prof. Dr. Robertas Alzbutas (Kaunas University of Technology, Technological Sciences, Informatics Engineering, T007)
Dissertation Defence Board of Informatics Engineering Science Field:
Prof. Dr. Tomas Blažauskas (Kaunas University of Technology, Technological Sciences, Informatics Engineering, T007) – chairperson
Prof. Dr. Nuno Manuel Garcia dos Santos (University of Lisbon, Portugal, Natural Sciences, Informatics, N009)
Prof. Dr. Nikolaj Goranin (Vilnius Gediminas Technical University, Technological Sciences, Informatics Engineering, T007)
Assoc. Prof. Dr. Darius Jegelevičius (Kaunas University of Technology, Technological Sciences, Electrical and Electronic Engineering, T001)
Prof. Dr. Tomas Krilavičius (Vytautas Magnus University, Technological Sciences, Informatics Engineering, T007)
Dissertation defence meeting will be at Rectorate Hall of Kaunas University of Technology (K. Donelaičio 73–402, Kaunas)
The doctoral dissertation is available at the library of Kaunas University of Technology (Gedimino g. 50, Kaunas) and on the internet: J. Sengupta el. dissertation.pdf
© J. Sengupta, 2025 “The text of the thesis may not be copied, distributed, published, made public, including by making it publicly available on computer networks (Internet), reproduced in any form or by any means, including, but not limited to, electronic, mechanical or other means. Pursuant to Article 25(1) of the Law on Copyright and Related Rights of the Republic of Lithuania, a person with a disability who has difficulties in reading a document of a thesis published on the Internet, and insofar as this is justified by a particular disability, shall request that the document be made available in an alternative form by e-mail to doktorantura@ktu.lt.”
Annotation: Subarachnoid Hemorrhage (SAH), a severe subtype of intracranial hemorrhage, requires timely detection and continuous monitoring to reduce complications and improve patient outcomes. This research presents a comprehensive Deep Learning (DL)-based system for SAH detection, segmentation, and severity assessment, developed through a multi-phase methodology designed to ensure clinical applicability and stability. A systematic literature review highlighted essential limitations in existing methods, including limited accuracy across diverse datasets, challenges in differentiating SAH from similar conditions, and high computational resources. To address these gaps, hybrid feature extraction and advanced segmentation techniques were introduced, significantly improving detection accuracy (from 95.45% to 99.36%) and segmentation performance (Modified region-growing method with similarity metric JC 0.94 and Super-pixel clustering with JC 0.92). Models such as OGRU and Bi-LSTM achieved superior classification accuracy (up to 99.62%) with reduced computational complexity (OGRU-CSO runtime of 33.28s). Furthermore, an SAH grading module was developed to enhance SAH severity assessment, demonstrating consistent performance under both noiseless and noisy image database, and the integration of datasets from both open-source repositories (RSNA) and Lithuanian hospitals ensured reliability and generalizability. By combining interdisciplinary insights from informatics engineering, radiology, and neurology, this work not only contributes to technological innovations but also establishes justifiable clinical relevance. Together, these achievements provide a clinically viable AI-driven solution for early SAH detection, precise segmentation, and severity monitoring, ultimately supporting improved diagnostic workflows.