Real-Time Counting People in Crowded Areas by Using Local Empirical Templates and Density Ratios

Abstract

SUMMARYIn this paper, a fast and automated method of counting pedestrians in crowded areas is proposed along with three contributions. We firstly propose Local Empirical Templates (LET), which are able to outline the foregrounds, typically made by single pedestrians in a scene. LET are extracted by clustering foregrounds of single pedestrians with similar features in silhouettes. This process is done automatically for unknown scenes. Secondly, comparing the size of group foreground made by a group of pedestrians to that of appropriate LET captured in the same image patch with the group foreground produces the density ratio. Because of the local scale normalization between sizes, the density ratio appears to have a bound closely related to the number of pedestrians who induce the group foreground. Finally, to extract the bounds of density ratios for groups of different number of pedestrians, we propose a 3D human models based simulation in which camera viewpoints and pedestrians' proximity are easily manipulated. We collect hundreds of typical occluded-people patterns with distinct degrees of human proximity and under a variety of camera viewpoints. Distributions of density ratios with respect to the number of pedestrians are built based on the computed density ratios of these patterns for extracting density ratio bounds. The simulation is performed in the offline learning phase to extract the bounds from the distributions, which are used to count pedestrians in online settings. We reveal that the bounds seem to be invariant to camera viewpoints and humans' proximity. The performance of our proposed method is evaluated with our collected videos and PETS 2009's datasets. For our collected videos with the resolution of 320x240, our method runs in real-time with good accuracy and frame rate of around 30 fps, and consumes a small amount of computing resources. For PETS 2009's datasets, our proposed method achieves competitive results with other methods tested on the same datasets [1], [2]. key words: local empirical templates, local density ratios, density ratio bounds, and people counting

The Estimation of Heights and Occupied Areas of Humans from Two Orthogonal Views for Fall Detection

“…The process of LET extraction can be considered as the initial setup of our proposed method. However, it is straightforward and can be done in an automated way (8) (9) .…”

“…(1) LET, by its nature, hold the information of the camera perspective. (2) LET extraction for unknown scenes is straightforward and can be done automatically since LET are merely the sizes of standing people in local image patches (8) (9) .…”

“…The sizes of foregrounds are extracted and kept in each cell buffer for clustering to generate an appropriate LET for the cell. The initial setup is straightforward and can be performed in an automated manner (8) (9) . It allows users without technical knowledge to customize the fall detection system for the different elderly and under various camera viewpoints.…”

“…For fair comparisons with existing methods of fall detection, the recommended methodology of performance evaluation is to conduct the experiments on the same dataset (9) . In this paper, we use the "Multi-view fall dataset" recently released by Auvinet et al (1) , which was adopted in the experiments of two latest studies (2) (3) .…”

“…In consideration of the above feature-state relationship, three states of humans fall into three separable regions of the proposed feature space, composing of normalized heights and normalized occupied areas, which can be classified by using support vector machines. In performance evaluation and comparison, a recommended methodology is to test the method on a common benchmark dataset (9) . This facilitates a fair comparison with other existing methods.…”