Person re-identification deals with the automatic tracking of individuals across cameras without overlapping field of view. Assisting security personnel in monitoring pedestrian behavior at wider areas with reduced costs is an important task. In person re-identification, we have a gallery set and the objective is to match a probe image with one of these individuals. Using image pairs, present in the training set, we can learn similarity models. One the main challenges in person re-identification is the higher inter-class similarities as opposed to intra-class similarities caused by similar dressing patterns and acquisition conditions (illumination, viewpoint, etc.). We have published works based on the combination of different color descriptors through a ranking aggregation strategy (CBRA) and the use of prototypes to indirectly handle with the camera transition problem.
Click here for a comprehensive list with results for several person re-identification datasets.