SwarMotion: A 3D Point Cloud Video Recording Tool for Classification Purposes

Abstract

Modern life is ever more reliant on computers being able to classify the world around them and computer vision is one of the ways computers do it. Nowadays, due to the advent of reliable and low-cost range sensors like Kinect which provide useful 3D data to feed prediction systems with a new dimension of useful information, computer vision is taking a new step with research demonstrating the potential that this kind of data has. However, very little research has been done using spatiotemporal Point Cloud data (PC-Videos). One reason might be the lack of datasets containing PC-Videos. In this paper, we propose SwarMotion, a multimodal recording tool focused on the acquisition of PC-Videos.