Crowd Counting and Profiling: Methodology and Evaluation

Abstract

Video imagery based crowd analysis for population profiling and density estimation in public spaces can be a highly effective tool for establishing global situational awareness. Different strategies such as counting by detection and counting by clustering have been proposed, and more recently counting by regression has also gained considerable interest due to its feasibility in handling relatively more crowded environments. However, the scenarios studied by existing regression-based techniques are rather diverse in terms of both evaluation data and experimental settings. It can be difficult to compare them in order to draw general conclusions on their effectiveness. In addition, contributions of individual components in the processing pipeline such as feature extraction and perspective normalisation remain unclear and less well studied. This study describes and compares the state-of-the-art methods for video imagery based crowd counting, and provides a systematic evaluation of different methods using the same protocol. Moreover, we evaluate critically each processing component to identify potential bottlenecks encountered by existing techniques. Extensive evaluation is conducted on three public scene datasets, including a new shopping centre environment with labelled ground truth for validation. Our study reveals new insights into solving the problem of crowd analysis for population profiling and density estimation, and considers open questions for future studies.

Advances in Crowd Analysis for Urban Applications Through Urban Event Detection

“…Loy et al [80] studied as well as compared the state-of-theart regression methods dealing with many co-linearity among features. They found that feature selections are important and depend on crowd scene.…”

Cascade Method for Image Processing Based People Detection and Counting

“…an indicator of fighting, rioting, violent protest, mass panic and excitement [4], [5];  business intelligence and behavioural economics applications; e.g. the distribution of costumers may be used for product placement, floor planning and staff management [6]. In addition, the overall crowd in a retail store may be monitored to assess store performance over time [7];  Transport applications; e.g.…”

Image Processing Based Ambient Context-Aware People Detection and Counting

“…The low-level features are used with the proposed system to describe the visual properties in the frames as texture, edge, shape, size and colour [5], [53]. Blob size histogram and edge orientation histogram are also used with the proposed system as intermediate features between low and high-level features.…”

“…They are the general description of a frame and have a strong relationship with the crowd size [5]. local binary pattern (LBP) and Gray-level co-occurrence matrix (GLCM) are usually used to find texture features [5], [54], [55].…”

“…Size features refer area, perimeter, perimeter-area ratio and blob count of the foreground pixels. Shape features include the perimeter orientation histogram [5].…”