Video anomaly detection based on locality sensitive hashing filters

Zhang, Ying; Lu, Huchuan; Zhang, Lihe; Ruan, Xiang; Sakai, S.

doi:10.1016/j.patcog.2015.11.018

Cited by 102 publications

(72 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…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%

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)

316

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

Section: Related Workmentioning

confidence: 99%

“…In general, existing abnormal event detection frameworks extract features at a local level [7,9,15,22,23,24,25,31,32,38], global (frame) level [21,26,27,28,33], or both [5,6,11]. All these approaches extract features without explicitly taking into account the objects of interest.…”

Section: Introductionmentioning

confidence: 99%

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)

316

166

View full text Add to dashboard Cite

show abstract

“…Hence, the approximation comes from the fact that not all training examples are considered as the possible nearest neighbor. People have investigated incorporating hashing-based techniques into distancebased anomaly detection systems [12,30,44].…”

Section: Fast Nearest Neighbors Searchmentioning

confidence: 99%

Fast Distance-Based Anomaly Detection in Images Using an Inception-Like Autoencoder

Sarafijanovic-Djukic

Davis

2019

Discovery Science

View full text Add to dashboard Cite

The goal of anomaly detection is to identify examples that deviate from normal or expected behavior. We tackle this problem for images. We consider a two-phase approach. First, using normal examples, a convolutional autoencoder (CAE) is trained to extract a low-dimensional representation of the images. Here, we propose a novel architectural choice when designing the CAE, an Inception-like CAE. It combines convolutional filters of different kernel sizes and it uses a Global Average Pooling (GAP) operation to extract the representations from the CAE's bottleneck layer. Second, we employ a distanced-based anomaly detector in the low-dimensional space of the learned representation for the images. However, instead of computing the exact distance, we compute an approximate distance using product quantization. This alleviates the high memory and prediction time costs of distance-based anomaly detectors. We compare our proposed approach to a number of baselines and state-of-the-art methods on four image datasets, and we find that our approach resulted in improved predictive performance.

show abstract

“…The model is based on singular spectrum analysis (SSA) [19,20] and locality-sensitive hashing (LSH) [21][22][23] for short-term forecasting. SSA is used for decomposing the original data into two components: the mean trend and the fluctuation component.…”

Section: Introductionmentioning

confidence: 99%

Short-term local prediction of wind speed and wind power based on singular spectrum analysis and locality-sensitive hashing

Liu

et al. 2018

J. Mod. Power Syst. Clean Energy

View full text Add to dashboard Cite

With the growing penetration of wind power in power systems, more accurate prediction of wind speed and wind power is required for real-time scheduling and operation. In this paper, a novel forecast model for shortterm prediction of wind speed and wind power is proposed, which is based on singular spectrum analysis (SSA) and locality-sensitive hashing (LSH). To deal with the impact of high volatility of the original time series, SSA is applied to decompose it into two components: the mean trend, which represents the mean tendency of the original time series, and the fluctuation component, which reveals the stochastic characteristics. Both components are reconstructed in a phase space to obtain mean trend segments and fluctuation component segments. After that, LSH is utilized to select similar segments of the mean trend segments, which are then employed in local forecasting, so that the accuracy and efficiency of prediction can be enhanced. Finally, support vector regression is adopted for prediction, where the training input is the synthesis of the similar mean trend segments and the corresponding fluctuation component segments. Simulation studies are conducted on wind speed and wind power time series from four databases, and the final results demonstrate that the proposed model is more accurate and stable in comparison with other models.

show abstract

Video anomaly detection based on locality sensitive hashing filters

Cited by 102 publications

References 26 publications

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

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

Fast Distance-Based Anomaly Detection in Images Using an Inception-Like Autoencoder

Short-term local prediction of wind speed and wind power based on singular spectrum analysis and locality-sensitive hashing

Contact Info

Product

Resources

About