Concept for Anonymous Re-Identification

This work presents a novel concept for object Re-Identification (ReID) based on multiple object detecting visual sensors with focus on data privacy and ethical usage of Artificial Intelligence (AI) enhanced deployments. By challenging the current state of the art solutions for ReID, the proposed concept is optimising individual privacy in terms of data protection and transportation. With attention to the laws, covered by the General Data Protection Regulation (GDPR) inside the European Union, AI driven applications especially with deployments in public places are regularly prohibited. With in this paper announced methods all personal data neither leaves the secured sensors, nor gets stored on any media at any point of time - which is a requirement for GDPR. Furthermore, most of the currently used methods are compared with temporal finite datasets in retrospect. Our vision is to design a GDPR compliant platform for near real time ReID that supports contemporaneous auditing and not after the fact.