General Introduction
In this talk, we give a brief summary on person re-identification, discussing the problem, existing methods [1-5], empirical studies and datasets.
Survey
Introduction
- Problem: definition, application, setting and difficulty.
Literature Review
- Finding Correspondence: re-identification by segmentation [1], by detection [2], by salience [3].
- Learning Transformation: re-identification by relative distance [4-5], by transfer learning [6], across views [7].
- Empirical Studies: common features, common distance measurement.
- Datasets: VIPeR, i-LIDS, ETHZ, CAVIAR4REID, EPFL&CAMPUS-Human.
Reference
- [1] Person Re-Identification by Symmetry-Driven Accumulation of Local Features. M. Farenzena and L. Bazzani and A. Perina and V. Murino and M. Cristani. In Proc. International Conference on Computer Vision and Pattern Recogntion (CVPR), 2010.
- [2] Custom Pictorial Structures for Re-identification. D.S. Cheng and M. Cristani and M.Stoppa and L. Bazzani and V. Murino. In Proc. British Machine Vision Conference (BMVC), 2011
- [3] Unsupervised Salience Learning for Person Re-identification. R. Zhao and W. Ouyang and X. Wang. In Proc. International Conference on Computer Vision and Pattern Recogntion (CVPR), 2013.
- [4] Person Re-Identification by Support Vector Ranking. B. Prosser and W. Zheng and S. Gong and T. Xiang. In Proc. British Machine Vision Conference (BMVC), 2010.
- [5] Person Re-identification by Probabilistic Relative Distance Comparison. W. Zheng and S. Gong and T. Xiang. In Proc. International Conference on Computer Vision and Pattern Recogntion (CVPR), 2011.
- [6] Human Re-identification with Transferred Metric Learning. W. Li and R. Zhao and X. Wang. In Proc. International Conference on Computer Vision and Pattern Recogntion (CVPR), 2012.
- [7] Locally Aligned Feature Transforms across Views. W. Li and X. Wang. In Proc. International Conference on Computer Vision and Pattern Recogntion (CVPR), 2013.