Gender and Body Mass Index Classification Using a Microsoft Kinect Sensor

This paper shows results on gender and body mass index (BMI) classification using anthropometric and gait information gathered from subjects using a Microsoft Kinect sensor. We show that it’s possible to obtain high accuracies when identifying gender, but that classifying BMI is a harder task. We also compare the performance of different machine learning algorithms and different combination of attributes, showing that Multilayer Perceptron outperforms Support Vector Machine and KNearest Neighbours in the proposed tasks.