Detecting Abnormal Events in Video Using Narrowed Normality Clusters

Ionescu, Radu Tudor; Smeureanu, Sorina; Popescu, Marius; Alexe, Bogdan

doi:10.1109/wacv.2019.00212

Cited by 85 publications

(50 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We can clearly see that the marked clusters are farthest from both normal distribution means, indicating that the containing points are indeed outliers. This assumption is also supported by the results obtained by Ionescu et al [59] in abnormal event detection in video, as they employ the same approach (based on k-means) to remove clusters of outlier motion and appearance samples. Nevertheless, our aim is to test out this assumption by quantifying the performance improvement of ShotgunWSD 2.0 over its previous version.…”

Section: ) Sense Embeddings After Outlier Removalsupporting

confidence: 59%

ShotgunWSD 2.0: An Improved Algorithm for Global Word Sense Disambiguation

Butnaru

Ionescu

2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

ShotgunWSD is a recent unsupervised and knowledge-based algorithm for global word sense disambiguation (WSD). The algorithm is inspired by the Shotgun sequencing technique, which is a broadly-used whole genome sequencing approach. ShotgunWSD performs WSD at the document level based on three phases. The first phase consists of applying a brute-force WSD algorithm on short context windows selected from the document, in order to generate a short list of likely sense configurations for each window. The second phase consists of assembling the local sense configurations into longer composite configurations by prefix and suffix matching. In the third phase, the resulting configurations are ranked by their length, and the sense of each word is chosen based on a majority voting scheme that considers only the top configurations in which the respective word appears. In this paper, we present an improved version (2.0) of ShotgunWSD which is based on a different approach for computing the relatedness score between two word senses, a step that stays at the core of building better local sense configurations. For each sense, we collect all the words from the corresponding WordNet synset, gloss and related synsets, into a sense bag. We embed the collected words from all the sense bags in the entire document into a vector space using a common word embedding framework. The word vectors are then clustered using k-means to form clusters of semantically related words. At this stage, we consider that clusters with fewer samples (with respect to a given threshold) represent outliers and we eliminate these clusters altogether. Words from the eliminated clusters are also removed from each and every sense bag. Finally, we compute the median of all the remaining word embeddings in a given sense bag to obtain a sense embedding for the corresponding word sense. We compare the improved ShotgunWSD algorithm (version 2.0) with its previous version (1.0) as well as several state-of-the-art unsupervised WSD algorithms on six benchmarks: SemEval 2007, Senseval-2, Senseval-3, SemEval 2013, SemEval 2015, and overall (unified). We demonstrate that ShotgunWSD 2.0 yields better performance than ShotgunWSD 1.0 and some other recent unsupervised or knowledge-based approaches. We also performed paired McNemar's significance tests, showing that the improvements of ShotgunWSD 2.0 over ShotgunWSD 1.0 are in most cases statistically significant, with a confidence interval of 0.01. INDEX TERMS Word sense disambiguation, shotgun sequencing, word embeddings, outlier removal.

show abstract

Section: ) Sense Embeddings After Outlier Removalsupporting

confidence: 59%

ShotgunWSD 2.0: An Improved Algorithm for Global Word Sense Disambiguation

Butnaru

Ionescu

2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

“…As evaluation metric, we employ the area under the curve (AUC) computed with regard to ground-truth annotations at the frame-level. The frame-level AUC metric used in most previous works [6,7,13,14,21,22,23,24,25,34,36] considers a frame as being a correct detection, if it contains at least one abnormal pixel. We adopt the same framelevel AUC definition as these previous works.…”

Section: Discussionmentioning

confidence: 99%

“…Abnormal event detection is commonly formalized as an outlier detection task [2,5,6,9,14,15,18,23,25,26,29,36,37,38,39], in which the main approach is to learn a model of familiarity from training videos and label the de-tected outliers as abnormal. Several abnormal event detection approaches [5,6,9,23,29] learn a dictionary of atoms representing normal events during training, then label the events not represented in the dictionary as abnormal.…”

Section: Related Workmentioning

confidence: 99%

“…Abnormal event detection in video has drawn a lot of attention in the past couple of years [7,11,12,13,14,21,22,24,27,28,31,33,34,36,37,38], perhaps because it is considered a challenging task due to the commonly accepted definition of abnormal events, which relies on context. An example that illustrates the importance of context is a scenario in which a truck is being driven on the street (normal event) versus a scenario in which a truck is being driven in a pedestrian area (abnormal event).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Object-Centric Auto-Encoders and Dummy Anomalies for Abnormal Event Detection in Video

Ionescu

Khan

Georgescu

et al. 2019

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

Self Cite

292

166

View full text Add to dashboard Cite

Abnormal event detection in video is a challenging vision problem. Most existing approaches formulate abnormal event detection as an outlier detection task, due to the scarcity of anomalous data during training. Because of the lack of prior information regarding abnormal events, these methods are not fully-equipped to differentiate between normal and abnormal events. In this work, we formalize abnormal event detection as a one-versus-rest binary classification problem. Our contribution is two-fold. First, we introduce an unsupervised feature learning framework based on object-centric convolutional auto-encoders to encode both motion and appearance information. Second, we propose a supervised classification approach based on clustering the training samples into normality clusters. A one-versus-rest abnormal event classifier is then employed to separate each normality cluster from the rest. For the purpose of training the classifier, the other clusters act as dummy anomalies. During inference, an object is labeled as abnormal if the highest classification score assigned by the one-versus-rest classifiers is negative. Comprehensive experiments are performed on four benchmarks: Avenue, ShanghaiTech, UCSD and UMN. Our approach provides superior results on all four data sets. On the large-scale ShanghaiTech data set, our method provides an absolute gain of 8.4% in terms of frame-level AUC compared to the state-of-the-art method [34].

show abstract

“…To organize a tight boundary to detect abnormal data, we come up with a second model refinement stage which can classify samples more precisely. Inspired by the one-class SVM which is proved to construct a tighter (or narrower) boundary gap [40], we adopt the one-class SVM to leave out the small percentage of samples who are mislabeled as normality in the first stage. To make more explicit explanation for its working principle, we visualize the tight frontier of one-class SVM on a peanut shape dataset in Figure 4.…”

Section: Proposed Methodsmentioning

confidence: 99%

An Efficient and Robust Unsupervised Anomaly Detection Method Using Ensemble Random Projection in Surveillance Videos

Zhu

Wang

et al. 2019

Sensors

View full text Add to dashboard Cite

Video anomaly detection is widely applied in modern society, which is achieved by sensors such as surveillance cameras. This paper learns anomalies by exploiting videos under the fully unsupervised setting. To avoid massive computation caused by back-prorogation in existing methods, we propose a novel efficient three-stage unsupervised anomaly detection method. In the first stage, we adopt random projection instead of autoencoder or its variants in previous works. Then we formulate the optimization goal as a least-square regression problem which has a closed-form solution, leading to less computational cost. The discriminative reconstruction losses of normal and abnormal events encourage us to roughly estimate normality that can be further sifted in the second stage with one-class support vector machine. In the third stage, to eliminate the instability caused by random parameter initializations, ensemble technology is performed to combine multiple anomaly detectors’ scores. To the best of our knowledge, it is the first time that unsupervised ensemble technology is introduced to video anomaly detection research. As demonstrated by the experimental results on several video anomaly detection benchmark datasets, our algorithm robustly surpasses the recent unsupervised methods and performs even better than some supervised approaches. In addition, we achieve comparable performance contrast with the state-of-the-art unsupervised method with much less running time, indicating the effectiveness, efficiency, and robustness of our proposed approach.

show abstract

Detecting Abnormal Events in Video Using Narrowed Normality Clusters

Cited by 85 publications

References 39 publications

ShotgunWSD 2.0: An Improved Algorithm for Global Word Sense Disambiguation

ShotgunWSD 2.0: An Improved Algorithm for Global Word Sense Disambiguation

Object-Centric Auto-Encoders and Dummy Anomalies for Abnormal Event Detection in Video

An Efficient and Robust Unsupervised Anomaly Detection Method Using Ensemble Random Projection in Surveillance Videos

Contact Info

Product

Resources

About