UHCTD: A Comprehensive Dataset for Camera Tampering Detection

Mantini, Pranav; Shah, Shishir K.

doi:10.1109/avss.2019.8909856

Cited by 8 publications

(9 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Exploiting this strategy, the experimental results were validated using k-fold validation (k = 3), choosing different sets of camera positions for each data partition. The external dataset investigated in the second series of experiments was the University of Houston Camera Tampering Detection Dataset (UHCTD) [39], proposed to test camera tampering detection methods. Tampering corresponds to an unauthorized or an accidental change in the view of a surveillance camera.…”

Section: Experimental Protocolmentioning

confidence: 99%

“…To maintain the same experimental protocol adopted for the other datasets investigated in this work, videos were sampled at 55 frames, with 224 × 224 spatial resolution. However, the data partitioning proposed by the UHCTD authors [39] was maintained, using the traditional hold-out validation strategy. Hence, after sampling, 30,252 samples of normal videos and 10,572 samples of anomalous videos were obtained to compose the training set.…”

Section: Experimental Protocolmentioning

confidence: 99%

See 1 more Smart Citation

Spatio-Temporal Deep Learning-Based Methods for Defect Detection: An Industrial Application Study Case

et al. 2021

View full text Add to dashboard Cite

Data-driven methods—particularly machine learning techniques—are expected to play a key role in the headway of Industry 4.0. One increasingly popular application in this context is when anomaly detection is employed to test manufactured goods in assembly lines. In this work, we compare supervised, semi/weakly-supervised, and unsupervised strategies to detect anomalous sequences in video samples which may be indicative of defective televisions assembled in a factory. We compare 3D autoencoders, convolutional neural networks, and generative adversarial networks (GANs) with data collected in a laboratory. Our methodology to simulate anomalies commonly found in TV devices is discussed in this paper. We also propose an approach to generate anomalous sequences similar to those produced by a defective device as part of our GAN approach. Our results show that autoencoders perform poorly when trained with only non-anomalous data—which is important because class imbalance in industrial applications is typically skewed towards the non-anomalous class. However, we show that fine-tuning the GAN is a feasible approach to overcome this problem, achieving results comparable to those of supervised methods.

show abstract

Section: Experimental Protocolmentioning

confidence: 99%

Section: Experimental Protocolmentioning

confidence: 99%

Spatio-Temporal Deep Learning-Based Methods for Defect Detection: An Industrial Application Study Case

et al. 2021

View full text Add to dashboard Cite

show abstract

“…On the other hand, the residual part E is expected to recover everything that is not static, namely the moving objects in the video. As mentioned before in Section 1, due to insufficient capabilities of existing robust SVD methods, an alternative principal component pursuit is used to retrieve the low-rank approximation L. In this section, we use the proposed rSVDdpd instead to obtain L. For the demonstration of rSVDdpd based video surveillance modeling and object extraction, we choose University of Houston Camera Tampering Detection Dataset (UHCTD) (Mantini and Shah, 2019a).…”

Section: Video Surveillance Background Modellingmentioning

confidence: 99%

“…With the problem of camera tampering detection (i.e., classification of tampered frames) in mind, Mantini and Shah (2019a) have compiled a comprehensive a large scale dataset called UHCTD (University of Houston Camera Tampering Detection Dataset). The dataset contains surveillance videos of over 288 hours ranging across 6 days from two cameras.…”

Section: Video Surveillance Background Modellingmentioning

confidence: 99%

“…Even though considerable effort has been given to solve this problem (see Garcia-Garcia, Bouwmans and Silva (2020) for details), various challenges such as presence of noisy frames, low frame rate of camera, change in illumination, multimodal backgrounds, presence of small moving objects, camera tampering still present major hindrances. Mantini and Shah (2019a) define tampering as a change in the view of a surveillance camera which may occur naturally or manually through unexpected or unauthorized actions. Reflection or glare of sunlight onto the camera lens, change in the camera view from daylight mode to night vision mode, defocusing of the camera lens, are some of the natural occurrences of tampering without any human intervention.…”

mentioning

confidence: 99%

See 1 more Smart Citation

A New Robust Scalable Singular Value Decomposition Algorithm for Video Surveillance Background Modelling

Roy¹,

Basu²,

Ghosh³

2021

Preprint

View full text Add to dashboard Cite

A basic algorithmic task in automated video surveillance is to separate background and foreground objects. Camera tampering, noisy videos, low frame rate, etc., pose difficulties in solving the problem. A general approach which classifies the tampered frames, and performs subsequent analysis on the remaining frames after discarding the tampered ones, results in loss of information. We propose a robust singular value decomposition (SVD) approach based on the density power divergence to perform background separation robustly even in the presence of tampered frames. We also provide theoretical results and perform simulations to validate the superiority of the proposed method over the few existing robust SVD methods. Finally, we indicate several other use-cases of the proposed method to show its general applicability to a large range of problems.

show abstract

Robust singular value decomposition with application to video surveillance background modelling

Roy,

Ghosh,

Basu

2024

Stat Comput

View full text Add to dashboard Cite

UHCTD: A Comprehensive Dataset for Camera Tampering Detection

Cited by 8 publications

References 22 publications

Spatio-Temporal Deep Learning-Based Methods for Defect Detection: An Industrial Application Study Case

Spatio-Temporal Deep Learning-Based Methods for Defect Detection: An Industrial Application Study Case

A New Robust Scalable Singular Value Decomposition Algorithm for Video Surveillance Background Modelling

Robust singular value decomposition with application to video surveillance background modelling

Contact Info

Product

Resources

About