Abnormal behavior detection using dominant sets

Alvar, Manuel; Torsello, Andrea; Sánchez-Miralles, Álvaro; Armingol, José María

doi:10.1007/s00138-014-0615-4

Cited by 29 publications

(11 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Unsupervised methods are easy and can be used to quickly detect abnormal behaviors [262], [263] • Supervised methods are useful to detect only particular behaviors, e.g., fighting [264], falling [13], and loitering [265]. • Supervised methods cannot detect ambiguous anomalies and is not practically feasible to learn all possible abnormal behaviors of humans [259].…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Vision-Based Human Detection Techniques: A Descriptive Review

Sumit

Mirjalili

2021

IEEE Access

View full text Add to dashboard Cite

Section: Resultsmentioning

confidence: 99%

“…• Semi-supervised methods are responsive to multiple parameters [266]. • Training time for unsupervised methods is relatively longer [262], [263] Crowd scene analysis…”

Section: Resultsmentioning

confidence: 99%

Vision-Based Human Detection Techniques: A Descriptive Review

Sumit

Mirjalili

2021

IEEE Access

View full text Add to dashboard Cite

“…Since Horn Shunck model [37] and Lucas Kanade model [44] were proposed in 1981, optical flow has been broadly applied to various fields, such as visual tracking [33], structure from motion (SFM) [22], motion segmentation [58,66], object recognition [26], video surveillance [1,3,30,49], visual odometry (VO) [16,55], and even video compression [20].…”

Section: Introductionmentioning

confidence: 99%

Real-time one-dimensional motion estimation and its application in computer vision

Yang

Gong

Tang

et al. 2015

Machine Vision and Applications

View full text Add to dashboard Cite

Traditional optical flow methods are designed to generate two-dimensional (2D) motion fields. However in some applications, only the motion field along a predefined line is needed, while the rest of the redundant 2D region (Cong et al., CVPR, 2009) can be ignored. In this paper, we propose a general framework for estimating one-dimensional (1D) motion field using the L 1 norm metric, which can be flexible for adding any constraints depending on the specific intention. Compared with 2D methods in model calculation, which often obtain a local optimal solution, our 1D version has the advantage of real-time global optimization and sub-pixel accuracy with a large displacement capacity. The synthesized experiments have justified the effectiveness of our 1D model. Two vision applications are investigated: (i) for crowd counting, we design a novel unified framework for both line of interest and region of interest counting using our 1D motion model to mosaic the dynamic blobs. Such a framework can adapt to new site without scene-specific learning and overcome overfitting. (ii) For visual odometry, we apply our 1D model to estimate the linear velocity of the robot in real time and obtain a comparable result with the benchmark Encoder. Experiments verified that our proposed model can produce satisfactory results in real time.

show abstract

“…Microscopic methods such as optical flow information [1] and movement [2] could be applied in the monitoring scene which could not be known in advance. The other is based on the classification of the lowlevel features, also known as the macroscopic method [3] [4]. But its accuracy rating largely depends on the precision of the edge detection.…”

Section: Introductionmentioning

confidence: 99%

Study on Abnormal Behavior in Elevator Based on Computer Vision

Shu¹,

Fu²,

Geng³

2014

Advanced Science and Technology Letters

View full text Add to dashboard Cite

As the violent issues occurred in the elevator are increasing, this paper discussed a detection algorithm of abnormal behavior. The characteristic differences were analyzed between normal behavior and violent behavior through the video image sequences captured in the elevator. The movement characteristics and kinetic energy of corner were composed a five dimensional feature vector as the identification basis of SVM to implement the violence detection. Comparing with the traditional threshold method and optical flow method, the algorithm was of low computation and higher recognition rate.

show abstract

Abnormal behavior detection using dominant sets

Cited by 29 publications

References 46 publications

Vision-Based Human Detection Techniques: A Descriptive Review

Vision-Based Human Detection Techniques: A Descriptive Review

Real-time one-dimensional motion estimation and its application in computer vision

Study on Abnormal Behavior in Elevator Based on Computer Vision

Contact Info

Product

Resources

About