Fig 1. The proposed framework for the image-to-video person re-identification.
More detail of this work can be found in [1].|
A reliable image-to-video person re-identification based on feature fusion |
| We formulate
person re-id problem as a classification-based information retrieval where a person appearance model is learned in the training phase and the identity of an interested person is determined by the probability that his/her
probe image belongs to the model. To learn the person appearance model, two features that are Kernel descriptor (KDES) and Convolution Neural
Network (CNN) are investigated. Then, three fusion schemes including early fusion, product rule and query-adaptive late fusions are proposed.
Extensive experiments have been conducted on two public benchmark datasets: CAVIAR4REID and RAID
Fig 1. The proposed framework for the image-to-video person re-identification. More detail of this work can be found in [1]. |
| Download: - Training and testing datasets used in the paper - KDES and CNN features - Code for extracing KDES, SVM classification, three fusion schemes [.zip] Please send email to Thi-Lan.Le@mica dot edu dot vn |
| References: |
| [1] NGUYEN Thuy-Binh, LE Thi Lan, NGUYEN Dinh-Duc, PHAM Dinh-Tan, A reliable image-to-video person re-identification based on feature fusion, 10th Asian Conference on Intelligent Information and Database Systems (ACIIDS), Springer - march 2018 |
| Home page |