Systematic Evaluation of Features From Pressure Sensors and Step Number in Gait for Age and Gender Recognition

Gait biometrics has grown sharply within decades and becomes an alternative approach for personal identification. However, the challenges of this application are difficult to record sufficient bioinformatics in gait for pattern recognition. Therefore, the study aimed to practically utilize the few features and steps for age and gender recognition. In total, 960 steps from twenty-four young and older participants were collected by using in-sole pressure mats. The center of the pressure trajectory was calculated, and thirty features were extracted according to the four phases within a step for the support vector machine (SVM). When only taking the three features, only five steps from each participant can achieve 97% accuracy for gender recognition, while 13 steps were required to achieve 95% for age recognition. The study revealed the features from the center of pressure for gait biometrics and demonstrated a small dataset for high accuracy recognition.