Abstract: Person re-identification is an automated video surveillance technique and has been extensively researched in recent years. The application of person re-identification covers the fields of security, robotics, multimedia, and forensics. The research problem which is often raised on the research topic of person re-identification is the feature representation which easily affected by occlusion (obstacles with other objects). In addition, local feature extraction through bounding boxes still contains the background image, so that it does not focus on parts of the human body. This research proposes a combination of methods between CNN, SVM classification, and semantic segmentation. Cumulative Matching Characteristics (CMC) and mean Average Precision (mAP) are evaluation metrics that will be used to measure re-identification performance..
Keywords:, Person re-identification; Feature extraction; CNN;
[1]. X. Y. Jing et al., "Super-Resolution Person Re-Identification with Semi-Coupled Low-Rank Discriminant Dictionary Learning," IEEE Trans. Image Process., vol. 26, no. 3, pp. 1363–1378, 2017, doi: 10.1109/TIP.2017.2651364.
[2]. L. Wu, Y. Wang, Z. Ge, Q. Hu, and X. Li, "Structured deep hashing with convolutional neural networks for fast person re-identification," Comput. Vis. Image Underst., vol. 167, no. December 2016, pp. 63–73, 2018, doi: 10.1016/j.cviu.2017.11.009.
[3]. A. Bedagkar-Gala and S. K. Shah, "A survey of approaches and trends in person re-identification," Image Vis. Comput., vol. 32, no. 4, pp. 270–286, 2014, doi: 10.1016/j.imavis.2014.02.001.
[4]. P. H. Tu et al., "An intelligent video framework for homeland protection," Unattended Ground, Sea, Air Sens. Technol. Appl. IX, vol. 6562, p. 65620C, 2007, doi: 10.1117/12.729215.
[5]. K. Wang, H. Wang, M. Liu, X. Xing, and T. Han, "Survey on person re-identification based on deep learning," CAAI Trans. Intell. Technol., vol. 3, no. 4, pp. 219–227, 2018, doi: 10.1049/trit.2018.1001
