Training Adversarial Discriminators for Cross-Channel Abnormal Event Detection in Crowds

Ravanbakhsh, Mahdyar; Sangineto, Enver; Nabi, Moin; Sebe, Nicu

doi:10.1109/wacv.2019.00206

Cited by 163 publications

(127 citation statements)

References 34 publications

Supporting

Mentioning

124

Contrasting

Unclassified

Order By: Relevance

“…Therefore, researchers (e.g., in [10]) have usually utilized pre-trained networks to extract features from the scenes, and the decision is surrendered to another module. (2) To train end-to-end models for this task, just recently [11][12][13][14] used generative adversarial networks (GANs) and adopted unsupervised methods learning the positive class (i.e., irregular events). In these methods, two networks (i.e., generator and discriminator) are trained.…”

Section: Introductionmentioning

confidence: 99%

“…We use an adversarial training scheme, similar to those used in generative adversarial networks (GANs) [15]. But in contrast to previous GAN-based models (e.g., [11,13,14,16]), we show how the two networks (I and D) can help each other to conduct the ultimate task of visual irregularity detection and localization. The two networks can be efficiently learned against each other, where I tries to inpaint the image such that D does not detect the whole generated image as irregularity.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

AVID: Adversarial Visual Irregularity Detection

Sabokrou

Pourreza²,

Fayyaz

et al. 2019

Computer Vision – ACCV 2018

View full text Add to dashboard Cite

Real-time detection of irregularities in visual data is very invaluable and useful in many prospective applications including surveillance, patient monitoring systems, etc. With the surge of deep learning methods in the recent years, researchers have tried a wide spectrum of methods for different applications. However, for the case of irregularity or anomaly detection in videos, training an end-to-end model is still an open challenge, since often irregularity is not well-defined and there are not enough irregular samples to use during training. In this paper, inspired by the success of generative adversarial networks (GANs) for training deep models in unsupervised or self-supervised settings, we propose an end-to-end deep network for detection and fine localization of irregularities in videos (and images). Our proposed architecture is composed of two networks, which are trained in competing with each other while collaborating to find the irregularity. One network works as a pixel-level irregularity Inpainter, and the other works as a patch-level Detector.After an adversarial self-supervised training, in which I tries to fool D into accepting its inpainted output as regular (normal), the two networks collaborate to detect and fine-segment the irregularity in any given testing video. Our results on three different datasets show that our method can outperform the state-of-the-art and fine-segment the irregularity.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

AVID: Adversarial Visual Irregularity Detection

Sabokrou

Pourreza²,

Fayyaz

et al. 2019

Computer Vision – ACCV 2018

View full text Add to dashboard Cite

show abstract

“…Therefore, GAN-based models employ different heuristics for the evaluation of novelty. For instance, in [38] a guided latent space search is exploited to infer it, whereas [35] directly queries the discriminator for a normality score.…”

Section: Related Workmentioning

confidence: 99%

Latent Space Autoregression for Novelty Detection

Abati

Porrello

Calderara

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

373

277

View full text Add to dashboard Cite

Novelty detection is commonly referred to as the discrimination of observations that do not conform to a learned model of regularity. Despite its importance in different application settings, designing a novelty detector is utterly complex due to the unpredictable nature of novelties and its inaccessibility during the training procedure, factors which expose the unsupervised nature of the problem. In our proposal, we design a general framework where we equip a deep autoencoder with a parametric density estimator that learns the probability distribution underlying its latent representations through an autoregressive procedure. We show that a maximum likelihood objective, optimized in conjunction with the reconstruction of normal samples, effectively acts as a regularizer for the task at hand, by minimizing the differential entropy of the distribution spanned by latent vectors. In addition to providing a very general formulation, extensive experiments of our model on publicly available datasets deliver on-par or superior performances if compared to state-of-the-art methods in one-class and video anomaly detection settings. Differently from prior works, our proposal does not make any assumption about the nature of the novelties, making our work readily applicable to diverse contexts.

show abstract

“…In both cases, inspired by [17], [20], [21], our architecture is composed by two fully-convolutional networks: the conditional generator G and the conditional discriminator D. The G network is the U-Net architecture [20], which is an encoder-decoder following with skip connections helping to preserve important local information. For D the PatchGAN discriminator [20], [22] is proposed, which is based on a "small" fully-convolutional discriminator. Hierarchy of cross-modal GANs: As described in Sec.I, the assumption is that the distribution of the normality patterns is under a high degree of diversity.…”

Section: B Private Layer Of Self-awarenessmentioning

confidence: 99%

Learning Multi-Modal Self-Awareness Models for Autonomous Vehicles from Human Driving

Ravanbakhsh

Baydoun

Campo

et al. 2018

2018 21st International Conference on Information Fusion (FUSION)

Self Cite

View full text Add to dashboard Cite

This paper presents a novel approach for learning self-awareness models for autonomous vehicles. Proposed technique is based on the availability of synchronized multi-sensor dynamic data related to different maneuvering tasks performed by a human operator. It is shown that different machine learning approaches can be used to first learn single modality models using coupled Dynamic Bayesian Networks; such models are then correlated at event level to discover contextual multimodal concepts. In the presented case, visual perception and localization are used as modalities. Cross-correlations among modalities in time is discovered from data and are described as probabilistic links connecting shared and private multi-modal DBNs at the event (discrete) level. Results are presented on experiments performed on an autonomous vehicle, highlighting potentiality of the proposed approach to allow anomaly detection and autonomous decision making based on learned self-awareness models.

show abstract

Training Adversarial Discriminators for Cross-Channel Abnormal Event Detection in Crowds

Cited by 163 publications

References 34 publications

AVID: Adversarial Visual Irregularity Detection

AVID: Adversarial Visual Irregularity Detection

Latent Space Autoregression for Novelty Detection

Learning Multi-Modal Self-Awareness Models for Autonomous Vehicles from Human Driving

Contact Info

Product

Resources

About