6th International Workshop on New Approaches for

Multidimensional Signal Processing

NAMSP 2025

Technical University of Sofia, Sofia, Bulgaria

July 24-26, 2025

---

Plenary Speakers

Prof. Dr. Szilvia Nagy, Department of Telecommunications at Széchenyi István University, Győr, Hungary

Title of Lecture:

Analysis of the current trends in computer aided brain lesion classification in MR images

Abstract: The brain is the central controlling organ of mammals, including humans, making the detection and classification of lesions in its tissue critically important. The advancement of computational intelligence methods over the past decade has significantly transformed brain lesion diagnosis, enabling more precise, efficient, and automated classification of tumors. Computer-aided diagnostic (CAD) systems play a crucial role in assisting radiologists by enhancing accuracy, reducing variability, and improving early detection. Among the most prominent approaches in the literature, neural networks, fuzzy logic-based models, and genetic algorithms have gained substantial attention due to their ability to process complex medical imaging data.
Deep learning, particularly convolutional neural networks (CNNs), is widely regarded as the most effective image-processing technique today, achieving remarkable accuracy in brain tumor classification. These models excel in feature extraction and pattern recognition, allowing them to outperform traditional machine learning approaches. Hybrid methods that integrate deep learning with fuzzy logic or genetic algorithms have also demonstrated promising results by refining classification boundaries and optimizing feature selection.
Techniques such as data augmentation, transfer learning, and ensemble modeling contribute to enhancing reliability and generalization across different datasets. However, applying these methods in the medical field requires particular caution, as biological variation is significantly greater than in conventional image processing tasks.
The performance evaluation of CAD systems relies on diverse metrics, with accuracy and the DICE similarity coefficient being the most widely used. Sensitivity, specificity, recall, and other domain-specific measures provide further insights into model robustness and clinical applicability.
Despite substantial progress, challenges remain, including the need for larger, well-annotated datasets, the ability to handle rare or highly heterogeneous tumors, and improved interpretability—especially crucial in medical applications. Addressing these challenges will be essential for the continued advancement of CAD methods, leading to more effective and accessible diagnostic solutions.

Biographical Notes:
Szilvia Nagy is a full professor at the Department of Telecommunications at Széchenyi István University, Győr. She graduated as an engineer physicist and earned her PhD in Physics from the Budapest University of Technology and Economics, later habilitating in Informatics at Széchenyi István University. Her research focuses on multidimensional signal processing, computational intelligence, and medical image analysis, with significant contributions to colorectal, liver, and brain lesion classification, AI-driven diagnostic methods, and mechanical and digital forensic signal processing. Her work extensively explores wavelet-based approaches and structural entropy for feature extraction and data analysis, contributing to advancements in signal processing and medical imaging.
She served as vice dean for research and international relations at the Faculty of Engineering Sciences at Széchenyi István University between 2011 and 2014 and was a working group vice leader at COST Action CA17124 DigForAsp (Digital Forensics: Evidence Analysis via Intelligent Systems and Practices) from 2018 to 2023. In recognition of her scientific contributions, she received the Hungarian Women in Science Excellence Award in 2016. She has published widely in leading scientific journals and conferences, shaping innovative methodologies at the intersection of artificial intelligence and biomedical engineering.