Optical flow techniques for crowd behavior analysis
摘要
This paper is concerned with optical flow techniques, crowd behavior and analysis from video sequences. Crowd appears in video sequences when the number of objects is incremented in the sequence with time. With large crowds, it is likely to encounter infection, violence, or accidents. That is why we adopt the trend of crowd analysis to determine the capacity of objects in a dedicated area with optical detection. Different methods have emerged in the literature for crowd analysis. These methods comprise deep learning and machine learning based on optical techniques. This paper presents a summary as well as a comparison of the methods adopted for this task. Furthermore, it explores the utilization of optical vehicle camera communication technologies in communication systems and shows their impact on crowd analysis. By illustrating various applications of these technologies combined with Artificial Intelligence (AI), we showcase their effectiveness in crowd analysis using optical techniques. Additionally, we provide insights into the state-of-the-art research on AI-based optical vehicle camera communication system.