Crowd Computing in Public Places

Computer Engineering

Crowd counting in public places plays a pivotal role in various domains including urban planning, security management, and event organization. This paper provides a comprehensive review and analysis of automated crowd counting techniques applied in public spaces. The study encompasses a wide array of methodologies ranging from traditional to state-of-the-art deep learning approaches. The review begins with an overview of the significance of crowd counting in public places, highlighting its importance in ensuring public safety, optimizing resource allocation, and enhancing crowd management strategies. Subsequently, it delves into the evolution of crowd counting techniques, tracing the progression from classical methods such as background subtraction and feature-based approaches to more recent advancements leveraging convolutional neural networks (CNNs), recurrent neural networks (RNNs), and their variants. Furthermore, the paper discusses various challenges encountered in crowd counting tasks, including occlusions, scale variations, and lighting conditions, along with the corresponding solutions proposed in literature. Additionally, it examines datasets commonly used for crowd counting research and evaluates their suitability for different scenarios. Moreover, the review critically assesses the performance metrics utilized to evaluate crowd counting algorithms and identifies the limitations of existing evaluation protocols. It also discusses the emerging trends and future directions in crowd counting research, such as the integration of multimodal data and the adoption of explainable AI techniques. Overall, this paper serves as a valuable resource for researchers, practitioners, and policymakers interested in understanding the state-of-the-art techniques, challenges, and future prospects of automated crowd counting in public spaces.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.