Top-push Video-based Person Re-identification.pdf


立即下载 NetworkAttachedStorage
2024-03-27
information re-id China Science matching problems. person video-based space-time University
665 KB

Top-push Video-based Person Re-identification
Jinjie You†,‡, Ancong Wu†,‡, Xiang Li†,‡, and Wei-Shi Zheng∗†,§,‡
†Intelligence Science and System Lab, Sun Yat-sen University, China
§School of Data and Computer Science, Sun Yat-sen University, China
‡Guangdong Provincial Key Laboratory of Computational Science, China
youjinjie9@gmail.com, wuancong@mail2.sysu.edu.cn
lixiang651@gmail.com, wszheng@ieee.org
Abstract
Most existing person re-identification (re-id) models fo-
cus on matching still person images across disjoint camer-
a views. Since only limited information can be exploited
from still images, it is hard (if not impossible) to overcome
the occlusion, pose and camera-view change, and lighting
variation problems. In comparison, video-based re-id meth-
ods can utilize extra space-time information, which con-
tains much more rich cues for matching to overcome the
mentioned problems. However, we find that when using
video-based representation, some inter-class differe


information/re-id/China/Science/matching/problems./person/video-based/space-time/University/ information/re-id/China/Science/matching/problems./person/video-based/space-time/University/
-1 条回复
登录 后才能参与评论
-->