Higuchi versus Katz fractal dimensions based features extraction method for epilepsy diagnosis using EEG signals

This paper presents a comparative analysis of Higuchi and Katz fractal dimensions algorithms (HFD and KFD), for EEG signal fractal analysis in epilepsy diagnosis, focusing on their effectiveness in feature extraction. For the first time in this work, a comparative study between three preprocessing approaches (Derivation, Empirical Mode Decomposition and Variational Mode Decomposition) is conducted using two algorithms HFD and KFD. This study contributes to the development of robust tools for epilepsy diagnosis which is proven with different experiments using the benchmark BONN database. Significant performance achievements evaluated through accuracy, sensitivity and selectivity metrics based on few features highlight the interest of our approach and reveal its applications for other purposes.