Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features
ECCV '08 Proceedings of the 10th European Conference on Computer Vision: Part I
Camera handoff and placement for automated tracking systems with multiple omnidirectional cameras
Computer Vision and Image Understanding
3D target tracking in distributed smart camera networks with in-network aggregation
Proceedings of the Fourth ACM/IEEE International Conference on Distributed Smart Cameras
Person re-identification in TV series using robust face recognition and user feedback
Multimedia Tools and Applications
Editor's Choice Article: A survey of approaches and trends in person re-identification
Image and Vision Computing
Hi-index | 0.00 |
This paper develops a concept of Panoramic Appearance Map (PAM) for performing person reidentification in a multi-camera setup. Each person is tracked in multiple cameras and the position on the floor plan is determined using triangulation. Using the geometry of the cameras and the person location, a panoramic map centered at the person’s location is created with the horizontal axis representing the azimuth angle and vertical axis representing the height. Each pixel in the map image gets color information from the cameras which can observe it. The maps between different tracks are compared using a distance measure based on weighted SSD in order to select the best match. Temporalintegration by registering multiple maps over the tracking period improves the matching performance. Experimental results of matching persons between two camera sets show the effectiveness of the approach.