Open Set Person Re-identification Framework on Closed Set Re-Id Systems

Person re-identification using analysis of photographs and video footages has a myriad of applications ranging from surveillance to human analytics. Techniques have been developed for closed-set re-identification of people in footages based on prior knowledge of the persons such as photographs, body features or face recognition models. However, except for face recognition, such methods are often incapable of handling the open-set scenario of identifying new persons that are not included in the current set of known persons. Additionally, face recognition is often not a practical solution for analysis of surveillance quality footages due to poor resolution and non-visibility of face. "Closed set to Open Set Person Re-Id Framework (C2OPR)" we propose in this paper, is a technique that could be used to extend a closed set person re-identification system to support identification of new persons. This is a generic approach, which can be used to extend any closed-set re-identification system to support open set re-identification. Through rigorous testing and improvements, we have reached accuracy in excess of 78% for identifying new persons using closed-set multi-view person re-identification systems.