Three different approaches based on 1D-LBP were proposed for GI The proposed methods are sensitive to new approach to noise High success rates were observed with the proposed approaches Gender identification (GI) is to determine the sex of the individual based on the characteristics that distinguish between male and female. In this study, three different feature extraction methods are proposed for gender identification by using signals obtained from accelerometers, magnetometers and gyroscope sensors installed in 5 different body parts of the individuals. 96.04%, 96.72% and 97.28% success rates were perceived with the recommended methods, respectively. In the study, attributes, motions, sensor port, sensor type that affect GI success were determined. The achievements of the proposed methods have been found to be more successful than the feature groups provided by the frequency and time domains of the proposed feature extraction methods, which are also compared with the success of the attribute groups derived from the same sensor signals from the time and frequency domains. Figure A. Gender identification system Purpose: In this study, three different feature extraction methods are proposed for gender identification as seen in the Fig. A. Theory and Methods: Feature extraction from signals is one of the most critical stages of GI. Because the success of GI depends on the attributes, different transformation methods have been applied to the signals obtained from sensors such as One Dimensional Local Binary Patterns (1D-LBPs), One Dimensional Robust Local Binary Patterns (1D-RLBPs) and Weighted One-Dimensional Robust Local Binary Patterns (W-1D-RLBPs). By using these attributes, different machine learning methods (SVM, RF, ANN, Knn) were elaborated for classification. Results: The features obtained by the proposed approaches are classified by different machine learning methods. The most successful classification method was observed as Knn. The success rates of 94.04%, 96.72% and 97.28% were observed with the recommended approaches. The W-1D-RLBP method was found to provide more effective attributes for GI than the other two methods. 12 different statistical features were calculated from the sensor signals. A high success rate of 97.06% was observed with the only entropy features in the trials to determine which feature was effective. In addition, grading was performed for each sensor type to determine the effect of accelerometers, gyroscopes and magnetometers on GI. The success rates were 91.33%, 89.21% and 89.43% respectively. It was found more appropriate to use these three types of sensors together. Conclusion: In all trials, the W-1D-RLBP approach was found to be more successful than the other 1D-LBP and 1D-RLBP approaches. As a result, it has been found that the proposed feature inference approaches provide effective attributes for GI. It is also considered that the proposed approaches can be applied to different signals.
