Flowsing: a multi-agent based offline ground-truth generator

An accurate ground-truth is critical in traffic monitoring and analysis related areas. Due to privacy concerns and technical difficulties, however, it is impossible to find certain public ground-truth with payload data included. It has become one of the most critical obstacles for traffic monitoring areas. On the other hand, ground-truth data made by hand is usually inefficient and hard to compare from one's work to another. In this paper we propose a novel architecture to automatically generate an accurate ground-truth including payload data. Based on the ground truth database obtained by our approach, researchers are able to analyze their proposed algorithms without the obstruction of untrustworthy data. Furthermore, we give a primary implementation of this architecture and make it available to the public.