Recognizing Threat through Object and Gesture Recognition by Applying Convolutional Neural Network and Adaptive Boosting Algorithm Using Microsoft Kinect for Xbox One
Roselito E. Tolentino, Janielle D. Antiporta, Carlo G. Coste, John Carlos H. Garro, Ariane Grace B. Hernandez, Mariela A. Pasamba · 2021
The study aimed to apply the Microsoft Kinect for Xbox One with the use of Convolution Neural Network and Adaptive Boosting to be able to recognize the hazardous object and threatening gestures. Microsoft Kinect for XBOX One and Visual Gesture Builder were used for obtaining the threatening object and gesture within the range of the main sensor. Whereas, the depth sensor is used for acquiring the gesture, and RGB-D Image Processing is for the use of distinguishing objects. Defined threat objects and threat gestures, the ones which are most common, are trained using supervised machine learning in order to recognize proper threat object and threat gesture. The system integrates the recognition of the two in order to define a threat recognition system.