arrow
Return

Message-passing framework for multi-camera worker tracking in construction

delete2025-10-17
delete0
PRE
AI
N
Nasrullah Khan
D
Dohyeong Kim
M
Minju Kim
D
Daeho Kim
D
Dongmin Lee *
DOI:10.1016/j.autcon.2025.106610delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• Addressed multi-camera tracking challenges in construction: similar attire, occlusion, and view transitions causing ID switches. • Proposed system using region-based re-ID for better features in occluded/similar-gear workers and message-passing for association. • Integrated detection per camera with localization, features, and motion for effective clustering and track generation. • Achieved IDF1 of 68.30 (controlled) and 85.10 (outdoor), MOTA 79.7/79.2, plus strong CAMPUS benchmark results. • Enables robust safety/productivity monitoring for construction sites, inspiring future occluded multi-view applications.
Keywords:
Multi-camera tracking (MCT)
Message passing
Re-identification
Construction safety monitoring
Computer vision
Tracking
Annotation tool

Journal

Automation in Construction cover
Automation in Construction
IF:
11.5
Papers:
6.3K
Citations:
4.2W

Organization

K
Korea University
Scholars:
3.6W
Papers: 3.8W
Citations: 4.4W
C
Chung-Ang University
Scholars:
1.2K
Papers: 549
Citations: 1.3W
K
Korea Institute of Economics Research
Scholars:
1
Papers: 1
Citations: 0
U
university of toronto
Scholars:
14.8W
Papers: 12.0W
Citations: 165
researcher View more organizations
Cited Papers

Cited Papers

errShare
errSave
errShare
errSave
Multi-Target Multi-Camera Tracking by Tracklet-to-Target Assignment
err2020-01-01
err70
PREAI
errHe, Yuhang; Wei, Xing; Hong, Xiaopeng; Shi, Weiwei; Gong, Yihong
errShare
errSave
Occluded Person Re-Identification
err2018-07-01
err0
errOAAI
errJiaxuan Zhuo; Zeyu Chen; Jianhuang Lai; Guangcong Wang
errShare
errSave
researcher View more