The Large-Scale Crowd Behavior Perception Based on Spatio-Temporal Viscous Fluid Field

Su, Hang; Yang, Hua; Zheng, Shibao; Yi, Fan; Wei, Sha Xin

doi:10.1109/tifs.2013.2277773

Cited by 53 publications

(38 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, LocalMotion 39 model is completely unsupervised, making no prior assumptions of what abnormal events may look like. But 10 false alarms were reported from UMN 16 dataset for this model 13 as the estimated°ows sometimes failed to exploit the spatial characteristic of crowd behavior. Besides, it assumes that abnormal events come up rarely, normal events occur often and will be got easily in the initial time.…”

Section: Introductionmentioning

confidence: 95%

“…Main gain of supervised learning-based approaches [8][9][10][11][12][13][14][15] clearly lies in the fact that normal and abnormal cases are known precisely. But it is essential to get all normal and abnormal models before detection.…”

Section: Introductionmentioning

confidence: 99%

“…Hence 13 false alarms were encountered from UMN 16 dataset. In ShearForce, 13 interaction force was estimated directly from shear force in viscous°uid by studying peculiarity of local motion patterns. If interaction force estimation can assess erroneously, false alarms are increased.…”

Section: Introductionmentioning

confidence: 99%

“…If interaction force estimation can assess erroneously, false alarms are increased. ShearForce 13 model was validated by randomly selecting 40% frames from each scenario for parameter optimizing during the learning procedure, and used the rest for testing. The reliable learning of unknown parameters using random samples was not precisely possible and thus eight false alarms were reported from UMN 16 dataset.…”

Section: Introductionmentioning

confidence: 99%

“…The use of eigenvalues to recognize crowd events is not new. 13,19 To recognize large- we track those points over the next frames using Pyramid-LK algorithm. [51][52][53] Images in Fig.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

An Eigenvalue Approach to Detect Flows and Events in Crowd Videos

Sharif

2017

J CIRCUIT SYST COMP

View full text Add to dashboard Cite

Analysis of°ows in crowd videos is a remarkable topic with practical implementations in many di®erent areas. In this paper, we present a wide overview of this topic along with our own approach to this problem. Our approach treats the di±culty of crowd°ow analysis by distinguishing single versus multiple°ows in a scene. Spatiotemporal features of two consecutive frames are extracted by optical°ows to create a three-dimensional tensor, which retains appearance and velocity information. Tensor's upper left minor matrix captures intensity structure. A normalized continuous rank-increase measure for each frame is calculated by a generalized interlacing property of the eigenvalues of these matrices. In essence, measure values put through the knowledge of existing°ows. Yet they do not go into e®ect desirably due to optical°ow estimation error and some other factors. A proper set of the degree of polynomial tting functions decodes their existence. But how can we estimate that set? Its detailed study is performed. Zero°ow, single°ow, multiple°ows, and interesting events are detected as frame basis using thresholds on the polynomial¯tting measure values. Plausible mean outputs of recall rate (88.9%), precision rate (86.7%), area under the receiver operating characteristic curve (98.9%), and accuracy (92.9%) reported from conducted experiments on PETS2009 and UMN benchmark datasets make clear and visible that our method gains high-quality results to detect°o ws and events in crowd videos in terms of both robustness and potency.

show abstract

Section: Introductionmentioning

confidence: 95%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…The use of eigenvalues to recognize crowd events is not new. 13,19 To recognize large- we track those points over the next frames using Pyramid-LK algorithm. [51][52][53] Images in Fig.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations