Research Article
Impostor Resilient Multimodal Metric Learning for Person Reidentification
Figure 1
Three Modals , , and in Image Space. Query and Gallery lie in modal , while one impostor for query lies in modal and the other in modal . Metric is learned using the impostor from modal , and is learned using the impostor from modal . Then, the obtained retrieval results of and are shown in Ranklist 1 and Ranklist 2, respectively. Correct Match is in green rectangle.