Old Is Gold: Redefining the Adversarially Learned One-Class Classifier Training Paradigm

Zaheer, Muhammad Zaigham; Lee, Jin Ha; Astrid, Marcella; Lee, Seung-Ik

doi:10.1109/cvpr42600.2020.01419

Cited by 108 publications

(91 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…as reported in literature [9,33,54,60] as well as observed in our experiments (baseline performances in Fig. 6), an AE can also often reconstruct anomalous examples.…”

Section: Introductionsupporting

confidence: 90%

“…Reconstruction Based Methods: A common way to tackle the one-class classification (OCC) problem is by utilizing autoencoders (AEs) which learn normal data representations by reconstructing the inputs [9,10,29,30,35,58]. However, since AEs can also wellreconstruct anomalous data [9,33,54,60], several researchers proposed memory based networks to limit reconstruction capability of AEs [9,35]. The idea is to use only the learned memory vectors for reconstruction, which helps in achieving higher reconstruction loss for anomalous inputs.…”

Section: Related Workmentioning

confidence: 99%

“…Moreover, there is no restriction on the types of anomaly events that may occur, making it cumbersome to collect sufficient anomaly examples. Therefore, anomaly detection is often seen as one-class classification (OCC) problem in which only normal data is used to train a novelty detection model [3,9,21,23,24,40,54].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Learning Not to Reconstruct Anomalies

Astrid¹,

Zaheer²,

Lee³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Video anomaly detection is often seen as one-class classification (OCC) problem due to the limited availability of anomaly examples. Typically, to tackle this problem, an autoencoder (AE) is trained to reconstruct the input with training set consisting only of normal data. At test time, the AE is then expected to well reconstruct the normal data while poorly reconstructing the anomalous data. However, several studies have shown that, even with only normal data training, AEs can often start reconstructing anomalies as well which depletes the anomaly detection performance. To mitigate this problem, we propose a novel methodology to train AEs with the objective of reconstructing only normal data, regardless of the input (i.e., normal or abnormal). Since no real anomalies are available in the OCC settings, the training is assisted by pseudo anomalies that are generated by manipulating normal data to simulate the out-of-normal-data distribution. We additionally propose two ways to generate pseudo anomalies: patch and skip frame based. Extensive experiments on three challenging video anomaly datasets demonstrate the effectiveness of our method in improving conventional AEs, achieving state-of-theart performance.

show abstract

“…as reported in literature [9,33,54,60] as well as observed in our experiments (baseline performances in Fig. 6), an AE can also often reconstruct anomalous examples.…”

Section: Introductionsupporting

confidence: 90%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Learning Not to Reconstruct Anomalies

Astrid¹,

Zaheer²,

Lee³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In spite of the growing interest in video anomaly detection [9, 10, 14-16, 19-21, 24, 29, 31, 36-38, 40, 43, 49, 51, 57, 58, 61, 63], which generated significant advances leading to impressive performance levels [14,15,18,24,29,53,56,57,61,63,64], the task remains very challenging. The difficulty of the task stems from two interdependent aspects: (i) the reliance on context of anomalies, and (ii) the lack of abnormal training data.…”

Section: Introductionmentioning

confidence: 99%

UBnormal: New Benchmark for Supervised Open-Set Video Anomaly Detection

Acsintoae¹,

Florescu²,

Mare³

et al. 2021

Preprint

View full text Add to dashboard Cite

Detecting abnormal events in video is commonly framed as a one-class classification task, where training videos contain only normal events, while test videos encompass both normal and abnormal events. In this scenario, anomaly detection is an open-set problem. However, some studies assimilate anomaly detection to action recognition. This is a closed-set scenario that fails to test the capability of systems at detecting new anomaly types. To this end, we propose UBnormal, a new supervised open-set benchmark composed of multiple virtual scenes for video anomaly detection. Unlike existing data sets, we introduce abnormal events annotated at the pixel level at training time, for the first time enabling the use of fully-supervised learning methods for abnormal event detection. To preserve the typical open-set formulation, we make sure to include disjoint sets of anomaly types in our training and test collections of videos. To our knowledge, UBnormal is the first video anomaly detection benchmark to allow a fair headto-head comparison between one-class open-set models and supervised closed-set models, as shown in our experiments. Moreover, we provide empirical evidence showing that UBnormal can enhance the performance of a state-of-the-art anomaly detection framework on two prominent data sets, Avenue and ShanghaiTech.

show abstract

“…According to the previous works, the anomaly detection methods can be generally divided into two types. Some anomaly detection methods are designed through reconstruction errors, which focus on modeling normal patterns in video sequences [ 3 – 5 , 7 , 8 , 10 , 11 ]. These methods learn the feature representation model of the normal pattern in the training phase and use the differences between the abnormal and normal samples to determine the final abnormal score of the test data during the testing phase, such as reconstruction errors or specific thresholds [ 7 – 14 ].…”

Section: Introductionmentioning

confidence: 99%

Abnormal Detection in Big Data Video with an Improved Autoencoder

Bian

Tang

2021

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

With the rapid growth of video surveillance data, there is an increasing demand for big data automatic anomaly detection of large-scale video data. The detection methods using reconstruction errors based on deep autoencoders have been widely discussed. However, sometimes the autoencoder could reconstruct the anomaly well and lead to missing detections. In order to solve this problem, this paper uses a memory module to enhance the autoencoder, which is called the memory-augmented autoencoder (Memory AE) method. Given the input, Memory AE first obtains the code from the encoder and then uses it as a query to retrieve the most relevant memory items for reconstruction. In the training phase, the memory content is updated and encouraged to represent prototype elements of normal data. In the test phase, the learned memory elements are fixed, and reconstruction is obtained from several selected memory records of normal data. So, the reconstruction will tend to be close to normal samples. Therefore, the reconstruction of abnormal errors will be strengthened for abnormal detection. The experimental results on two public video anomaly detection datasets, i.e., Avenue dataset and ShanghaiTech dataset, prove the effectiveness of the proposed method.

show abstract

Old Is Gold: Redefining the Adversarially Learned One-Class Classifier Training Paradigm

Cited by 108 publications

References 40 publications

Learning Not to Reconstruct Anomalies

Learning Not to Reconstruct Anomalies

UBnormal: New Benchmark for Supervised Open-Set Video Anomaly Detection

Abnormal Detection in Big Data Video with an Improved Autoencoder

Contact Info

Product

Resources

About