The Journal of Building Design and Environment (JBDE), the official academic journal of the International Society for Smart Construction and Production (SCP), has recently published a study entitled "Lightweight head-based multi-object tracking in crowded built environments." Conducted by Yue Cui, Zijie Gao, and Kailai Sun, the research presents an efficient computer vision framework for tracking multiple individuals in crowded environments while reducing computational complexity.

Research Highlights

Accurate multi-object tracking is essential for intelligent building management, crowd monitoring, and construction site safety. However, conventional tracking systems often face challenges such as frequent occlusion, complex backgrounds, and high computational demands.

To address these issues, the researchers developed a lightweight head-based tracking framework integrating three key techniques:

* Head-Based Detection: Improves tracking in crowded environments where full-body detection is often affected by occlusion.

* Structured Channel Pruning: Reduces redundant computations and model complexity while maintaining essential detection capabilities.

* Quantization-Aware Training: Enhances computational efficiency and facilitates deployment on resource-constrained edge devices.

The study also evaluates the framework under challenging visual conditions, demonstrating its potential for practical monitoring applications.

Applications and Significance

The proposed framework offers promising opportunities for intelligent building operations, pedestrian flow analysis, crowd management, and construction site monitoring. Its lightweight architecture supports the development of efficient, edge-based artificial intelligence solutions for smart construction and urban infrastructure.

As SCP’s official academic journal, JBDE provides an international platform for disseminating interdisciplinary research in building design, environmental performance, intelligent construction, and emerging digital technologies. This publication further highlights the integration of artificial intelligence and building science in advancing smarter and more sustainable built environments.

Publication Information

Article: Lightweight head-based multi-object tracking in crowded built environments

Authors: Yue Cui, Zijie Gao, and Kailai Sun

Journal: Journal of Building Design and Environment (JBDE)

Published: October 16, 2026

DOI: https://doi.org/10.70401/jbde.2026.0046