Author, Institution: Sarmad Maqsood, Kaunas University of Technology
Science area, field of science: Technological Sciences, Informatics Engineering, T007
Scientific Supervisor: Prof. Dr. Robertas Damaševičius (Kaunas University of Technology, Technological Sciences, Informatics Engineering, T007)
Dissertation Defense Board of Informatics Engineering Science Field:
Prof. Dr. Agnė Paulauskaitė-Tarasevičienė (Kaunas University of Technology, Natural Sciences, Informatics, N009) – chairperson
Senior Researcher Dr. Jolita Bernatavičienė (Vilnius University, Technological Sciences, Informatics Engineering, T007)
Prof. Dr. Arnas Kačeniauskas (Vilnius Gediminas Technical University, Technological Sciences, Informatics Engineering, T007)
Prof. Dr. Simona Ramanauskaitė (Vilnius Gediminas Technical University, Technological Sciences, Informatics Engineering, T007)
Dr. Ihsan Ullah (University of Galway, Ireland, Technological Sciences, Informatics Engineering, T007)
Dissertation defense 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)
Annotation: The advent of smart healthcare demonstrates the rapid advancements in the information technology sector. This paradigm leverages cutting-edge technologies, such as artificial intelligence (AI) to revolutionize medical practices for enhanced efficiency, reliability, and personalization. Specifically, in cancer research, where accurate and timely diagnosis is critical, computer-aided diagnosis (CAD) systems are crucial for assisting healthcare professionals in diagnosing lesions through detection, segmentation, and classification of medical imagery. Despite the progress, challenges such as low-contrast lesions, imbalanced datasets, overfitting in convolutional neural network (CNN) models, memory complexity, and redundant feature extraction persist. This dissertation proposed an integrated CAD model using a deep learning (DL) framework to address these issues in lesion detection, segmentation, and classification. This dissertation designed an AI system for the classification of different diseases using a hybrid method. A custom CNN architecture is developed for segmenting the lesion regions, followed by the modification and training of several pre-trained CNN models via transfer learning (TL) on the segmented images. Deep feature vectors extracted from these models are then combined using a convolutional sparse image decomposition fusion method. Ultimately, feature selection techniques are employed to identify and utilize the most informative features for classification, promising a more effective and accurate diagnostic process.
May 8 d. 13:00
Rectorate Hall at Kaunas University of Technology (K. Donelaičio 73-402, Kaunas)
Įtraukti į iCal
