Real-Time On-Board Counting System in Crowded Scenes

Abstract

ABSTRACT : Crowd counting is a useful tool for situational awareness in public spaces. Automated crowd counting with videos and images is an interesting but difficult task that has attracted a lot of interest in computer vision. Different deep learning techniques have been developed recently to reach cutting-edge performance. Numerous features of the techniques that have been evolved throughout time cover model architecture, learning paradigm, computing complexity, input pipeline, accuracy gains, etc. Many researchers are devoting to crowd counting, and many excellent works of literature and works have spurted out. These works usually aim to be be helpful for the development of crowd counting. However, the question we should consider is why and how they are effective for this task. In this paper, we have surveyed many works to comprehensively and systematically study the crowd counting models, mainly CNNbased density map estimation methods. This thesis is aimed to categorize, analyze as well as provide the latest development and performance evolution in crowd counting using different deep learning techniques and methods that are published in journals and conferences over the past five years.