Identity Recognition through Anthropometric Feature Analysis of mmBody Data

With the development of biometrics, human feature recognition has become the focus of research. Traditional methods such as facial recognition and fingerprint recognition face challenges in low-light and non-contact environments. This study introduces a method for identification using human features extracted from mmBody data. The study focuses on extracting features from the upper body, such as head and neck length, upper body length and arm length. These measurements are combined to form feature sets. The validity of these features was verified experimentally and the feature pattern with the highest recognition rate was identified. For model training, algorithms such as feature matching and converter models were explored and evaluated through crossvalidation to ensure the robustness of the approach. The study utilized the mmBody dataset for experiments. The method aims to overcome the limitations of traditional biometric methods and provide a reliable alternative for authentication in challenging environments such as low-light conditions or where contactless identification is required.