重心・重量データを計測可能な床型デバイスを用いた歩容取得による性別識別

Recently, biometric methods based on gait, a characteristic of human walking, have garnered significant attention. In this study, we propose a method for identifying the gender of a walker using a floor-based device capable of measuring the center of gravity and weight data during walking. Compared to image sensor-based approaches, center-of-gravity and weight data offer the advantages of not infringing on privacy and being less susceptible to disturbances such as lighting, clothing, and hairstyle. The floor-type device is constructed from aluminum and equipped with load cells at eight locations to measure weight during walking and calculate the center of gravity’s position. Time-series data were collected from 31 subjects, and frequency features were extracted using a Fast Fourier Transform. Group-stratified cross-validation revealed that gender classification was achieved with an accuracy of approximately 76.0 % when subjects walked under conditions similar to daily life, such as carrying a personal item. Additionally, an attempt was made to classify gender using only center-of-gravity sway features without weight data; however, the results were close to the chance rate, indicating that weight information is crucial at this stage.