An introduction to support Vector Machines: and other kernel-based learning methods
An introduction to support Vector Machines: and other kernel-based learning methods
Mean Shift: A Robust Approach Toward Feature Space Analysis
IEEE Transactions on Pattern Analysis and Machine Intelligence
Person Reidentification Using Spatiotemporal Appearance
CVPR '06 Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Volume 2
The Dissimilarity Representation for Pattern Recognition: Foundations And Applications (Machine Perception and Artificial Intelligence)
Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features
ECCV '08 Proceedings of the 10th European Conference on Computer Vision: Part I
Video Sequences Association for People Re-identification across Multiple Non-overlapping Cameras
ICIAP '09 Proceedings of the 15th International Conference on Image Analysis and Processing
Person Re-identification Using Haar-based and DCD-based Signature
AVSS '10 Proceedings of the 2010 7th IEEE International Conference on Advanced Video and Signal Based Surveillance
Person Re-identification Using Spatial Covariance Regions of Human Body Parts
AVSS '10 Proceedings of the 2010 7th IEEE International Conference on Advanced Video and Signal Based Surveillance
A multiple component matching framework for person re-identification
ICIAP'11 Proceedings of the 16th international conference on Image analysis and processing - Volume Part II
Fast person re-identification based on dissimilarity representations
Pattern Recognition Letters
Hi-index | 0.00 |
Among the possible applications of computer vision to video-surveillance, person re-identification over a network of camera sensors, using cues related to clothing appearance, is gaining much interest. Re-identification techniques can be used for various tasks, e.g., online tracking of a person, and off-line retrieval of all video sequences containing an individual of interest, whose image is given as a query. Recently, some authors proposed to exploit clothing appearance descriptors also to retrieve video sequences of individuals that match a textual description of clothing (e.g., "person wearing a black t-shirt and white trousers"), instead of an image. We name this task "appearance-based people search". This functionality can be useful, e.g., in forensics investigations, where a textual description can be provided by a witness. In this paper, we present and experimentally evaluate a general method to perform both person re-identification and people search, using any given descriptor of clothing appearance that exploits widely used multiple part/multiple component representations. It is based on turning the considered appearance descriptor into a dissimilarity-based one, through a framework we previously proposed for speeding up person re-identification methods. Our approach allows one to deploy systems able to perform both tasks with the same pipeline and processing stages for constructing descriptors.