Crowd flux analysis and abnormal event detection in unstructured and structured scenes

• Published in: Multimedia tools and applications Vol. 72; no. 3; pp. 3001 - 3029
• Main Author: Nam, Yunyoung
• Format: Journal Article
• Language: English
• Published: Boston Springer US 01.10.2014; Springer; Springer Nature B.V
• Subjects: Applied sciences; Artificial intelligence; Cameras; Computer Communication Networks; Computer Science; Computer science; control theory; systems; Crowds; Data processing. List processing. Character string processing; Data Structures and Information Theory; Escalators; Exact sciences and technology; False alarms; Field of view; Flux; Highways; Memory organisation. Data processing; Motion detectors; Multimedia; Multimedia communications; Multimedia Information Systems; Pattern recognition. Digital image processing. Computational geometry; Principal components analysis; Representations; Roads & highways; Software; Special Purpose and Application-Based Systems; Studies; Thresholds; Vision systems; United States > US; Spatio-temporal; Entropy; Moving objects; Crowded scene; Abnormality; Computer vision; Event detection; Visual field; Grid; Video signal; Crowd; Crowded scene; False alarm rate; Scene analysis; Object oriented; Freeway; Receiver operating characteristic curves; Anomaly; Moving body; Optical flow
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-013-1593-7

This paper presents a spatio-temporal grid-based framework to deal with the complexity of structured and unstructured motion flows that can effectively group optical flows in the field of view into crowds. This approach utilizes motion flows of the features based on a grid in a scene. In order to detect abnormal events in crowded scenes, the proposed method measures motion features including the speed and direction of moving objects based on a spatio-temporal grid-based approach for flow representation. Experiments have been conducted on several different videos in three domains that are crosswalks, escalators, and highways. To evaluate and compare the performance of our method to other methods, ROC curves are plotted which take into consideration both detection rate and false alarm rate for multiple threshold values.