Akademik Veri Yönetim Sistemi
International Conference on Scientific and Innovative Studies ICSIS 2023, Konya, Türkiye, 18 - 20 Nisan 2023, cilt.1, sa.29801931, ss.118-120, (Özet Bildiri)
Hand-movement features, or characteristics of a person's biometric pattern, can reveal a lot of information and so the number of features included in hand-movement is very large. Therefore, extracting discriminative hand-movement features can help in identifying biometric movements with high accuracy, tracking changes in patterns over time [1-5]. Studies have been conducted to define the functions of variability, which have significant roles in understanding [6-7]. The present study investigates the importance of distinctive hand-movement features. The question is which hand-movement features are most likely discriminative in healthy individuals and do they correspond to considerable description of human characteristics? Carnegie Mellon University's motion capture library [8] is utilized containing 32 subjects’ gait sequences, and each sequence contains at least 300-500 frames with a length of 3 to 6 seconds. These gait videos are captured using 12-16 retro-reflective markers. After pre-processing, each movement signals are decomposed into sub-signals utilizing the feature extraction methods wavelet transform-based decompositions (WT-based decomposition), empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD), and variational mode decomposition (VMD). Due to the fact that the size of feature data is numerous, this approach is proposed as a way to reduce calculation costs and increase accuracy through the use of mRMR. mRMR relies on mutual information to determine the variables that are most discriminatory after measuring similarity between them