Pedestrian Detection and Tracking for Counting Applications in Crowded Situations

Abstract

This paper describes a vision based pedestrian detection and tracking system which is able to count people in very crowded situations like escalator entrances in underground stations. The proposed system uses motion to compute regions of interest and prediction of movements, extracts shape information from the video frames to detect individuals, and applies texture features to recognize people. A search strategy creates trajectories and new pedestrian hypotheses and then filters and combines those into accurate counting events. We show that counting accuracies up to 98 % can be achieved.

Automatic Detection of Unexpected Events in Dense Areas for Videosurveillance Applications

“…Kratz & Nishimo (2010) distinguish people inside a crowd using both color histogram and global mouvement model. In tracking, the assumption is that the shape of persons does not vary much at the scale of and individual in a crowd, and that physical points lying on a person move in the same way (same trajectory and same speed) (Brostow & Cipolla (2006), Rabaud & Belongie (2006), Sidla & Lypetskyy (2006) and Hu et al (2006)). Tracking algorithms are widely used to recover people trajectories.…”

“…Learning paths enables the detection of abnormal trajectories (Junejo & Foroosh (2007), Hu et al (2006), Saleemi et al (2008)), or infering people interaction (Blunsden et al (2007), Oliver et al (2000)). The analysis of trajectories is also used in intrusion detection applications where crossing a virtual line raises an alarm or increase a counter ( Rabaud & Belongie (2006) or Sidla & Lypetskyy (2006)). Local approaches also tackles the problem of posture recognition in crowded area (Zhao & Nevatia (2004), Pham et al (2007)).…”

“…Methods differ in both the human appearance models (descriptors) used for segmenting people and the way datas are associated. In Zhao & Nevatia (2004), Sidla & Lypetskyy (2006) and Pham et al (2007), individuals are segmented using several ellipses or rectangles to represent body parts or the omega shape to model both the head and the shoulders. Yu & Medioni (2009) reinforce people tracking with occlusion by adding information on the appearance of the persons before they are occluded and an assumption on the speed continuity of tracked blobs.…”

Estimation of Human Traffic Flow Using Feature-based Regression in the Spatiotemporal Domain

“…Detection of whole human bodies often suffer due to occlusions in dense crowds. To resolve miss-detection due to occlusions, only upper body shapes (Sidla et al, 2006) or contours around heads (Yuk et al, 2006) may be used in detection. Part-based detection methods have been studied extensively (Wu and Nevatia, 2005;Lin et al, 2007) to improve detection performance in dense crowds.…”

An Unsupervised Learning Based Approach for Unexpected Event Detection

“…These points are gathered either with a bayesian framework as in (Brostow and Cipolla, 2006) and (Li and Ai, 2007), or with the RANSAC algorithm (Rabaud and Belongie, 2006). Hybrid methods combine both approaches to segment the crowd, in order to make the system more robust as in (Sidla and Lypetskyy, 2006). In the case of traffic monitoring, Hu (Hu et al, 2006) (with the additional hypothesis that an entity cannot partially hide another one), gathers features (spatial ones or trajectories) with Kmeans algorithm.…”