Deep-anomaly: Fully convolutional neural network for fast anomaly detection in crowded scenes

By Mohammad Sabokrou, Mohsen Fayyaz, Mahmood Fathy, Zahra Moayed, Reinhard Klette. Computer Vision and Image Understanding, 2016

470 citations recorded by the source corpus.

The detection of abnormal behaviours in crowded scenes has to deal with many challenges. This paper presents an efficient method for detection and localization of anomalies in videos. Using fully convolutional neural networks (FCNs) and temporal data, a pre-trained supervised FCN is transferred into an unsupervised FCN ensuring the detection of (global) anomalies in scenes. High performance in terms of speed and accuracy is achieved by investigating the cascaded detection as a result of reducing computation complexities. This FCN-based architecture addresses two main tasks, feature representation and cascaded outlier detection. Experimental results on two benchmarks suggest that detection and localization of the proposed method outperforms existing methods in terms of accuracy.

Subfields: Artificial Intelligence, Computer Networks and Communications, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/e6a8ca84b2bb90e2861fe0d1f05b54cae8321dd5

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Networks and Communications

In Computer Vision and Pattern Recognition

Citations in the corpus

Uninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings (2019, background): “In existing work for anomaly detection with pretrained networks, feature extractors only output single feature vectors for patchsized inputs or spatially heavily downsampled feature maps [20, 26].”
Multiresolution Knowledge Distillation for Anomaly Detection (2020, methodology): “[35] attempts to solve this by extracting lots of different patches from normal images.”
Anomaly Detection in Video via Self-Supervised and Multi-Task Learning (2020, methodology): “Without being able to employ standard supervision, researchers have proposed alternative approaches ranging from distance-based [17, 19, 37, 38, 40, 44, 45, 46, 47, 50, 52, 59] and reconstruction-based strategies [5, 13, 14, 27, 29, 31, 34, 36, 41, 51, 53] to probabilistic [1, 2, 4, 12, 16, 21, 32, ”
Self-Supervised Predictive Convolutional Attentive Block for Anomaly Detection (2021, methodology): “Anomaly detection methods can be classified into: dictionary learning methods [7–9, 14, 40, 55], change detection frameworks [11, 26, 38, 48], probability-based methods [1, 2, 16, 23, 29, 44, 45, 56, 61, 74], distance-based models [6, 10, 25, 27, 50, 51,53,57,58,63,65,68,70], and reconstruction-base”
Abnormal event detection in videos using generative adversarial nets (2017, background): “A deep learning-based approach has been investigated also in [15, 16].”
Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flows (2020, background): “[25] models the anomaly-free feature distribution as an unimodal Gaussian distribution.”
Student-Teacher Feature Pyramid Matching for Anomaly Detection (2021, methodology): “To avoid cropping image patches and accelerate feature extraction, Sabokrou et al. [33] build descriptors from early feature maps of a pre-trained fully convolutional network (FCN) and adopt a unimodal Gaussian distribution to ﬁt feature vectors of the anomaly-free images.”
Adversarially Learned One-Class Classifier for Novelty Detection (2018, background): “Some efforts have been made, in recent years, to beneﬁt from deep features in learning one-class classiﬁers [48, 39, 33, 20, 40, 38], few of which could train an end-to-end feature learning and classiﬁcation model.”
Lucid: A Practical, Lightweight Deep Learning Solution for DDoS Attack Detection (2020, background): “Convolutional Neural Networks (CNNs), a specific DL technique, have grown in popularity in recent times leading to major innovations in computer vision [6]–[8] and Natural”