SDK per a la representació d’objectes en 3D

Abstract

[en] Creation of an SDK where the main goals are camera and frames information gathering and classification, the main frame of the capture detection and video creation applied to the food domain made by the frames captured using algorithms implemented in Android, is the main topic of this work. We have studied how to capture the correct and necessary images to build a representation of food based on the center of the scene and the angle created between the first frame and the center of the scene. To achieve this goal we have implemented several algorithms to detect, compute and update the center of the scene dynamically and also compute the angles and classify them into ranges in real time to achieve an optimal 3D representation. Furthermore, the Kalman filter has been implemented to provide more robustness to the data acquisition, preventing internal errors in obtaining information from the captured frame camera. Also, since the capture of frames is done dynamically, it allows us to improve motion errors made unconsciously by the user.