OCGAN: One-Class Novelty Detection Using GANs With Constrained Latent Representations

By Pramuditha Perera, Ramesh Nallapati, Bing Xiang. 2019

596 citations recorded by the source corpus.

We present a novel model called OCGAN for the classical problem of one-class novelty detection, where, given a set of examples from a particular class, the goal is to determine if a query example is from the same class. Our solution is based on learning latent representations of in-class examples using a de-noising auto-encoder network. The key contribution of our work is our proposal to explicitly constrain the latent space to exclusively represent the given class. In order to accomplish this goal, firstly, we force the latent space to have bounded support by introducing a tanh activation in the encoder's output layer. Secondly, using a discriminator in the latent space that is trained adversarially, we ensure that encoded representations of in-class examples resemble uniform random samples drawn from the same bounded space. Thirdly, using a second adversarial discriminator in the input space, we ensure all randomly drawn latent samples generate examples that look real. Finally, we introduce a gradient-descent based sampling technique that explores points in the latent space that generate potential out-of-class examples, which are fed back to the network to further train it to generate in-class examples from those points. The effectiveness of the proposed method is measured across four publicly available datasets using two one-class novelty detection protocols where we achieve state-of-the-art results.

Subfields: Artificial Intelligence, Epidemiology, Molecular Biology

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 Epidemiology

In Molecular Biology

Citations in the corpus

Continual Learning by Using Information of Each Class Holistically (2021, background): “Since PCL learns one class at a time, it is related to one-class learning (Sch¨ olkopf et al. 2001; Tax and Duin 2004; Hu et al. 2020; Perera, Nallapati, and Xiang 2019; Ruff et al. 2018).”
A Wholistic View of Continual Learning with Deep Neural Networks: Forgotten Lessons and the Bridge to Active and Open World Learning (2020, background): “For now, we will start with a typical categorization of existing deep continual works into the three categories of regularization, rehearsal and architectural approaches, in consistency ∙ LWF (LiandHoiem,2016) ∙ EBLL (Rannenetal.,2017) ∙ Pseudorehearsal (Robins,1995) ∙ Pseudo-recurrentnets (French,1”
PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localization (2020, unknown): “Although reconstruction-based methods are very intuitive and interpretable, their performance is limited by the fact that AE can sometimes yield good reconstruction results for anomalous images too [21].”
Multiple Class Novelty Detection Under Data Distribution Shift (2020, unknown): “Recent works such as OC-GAN [38] and non-adversarial generative method [7] consider a speciﬁc case where it is assumed that there is only one category available in the dataset.”
A Unifying Review of Deep and Shallow Anomaly Detection (2020, background): “Further variants of deep one-class classification include multimodal [145] or time-series extensions [338] and methods that employ adversarial learning [138], [141], [339] or transfer learning [139], [142].”
Hybrid Models for Open Set Recognition (2020, background): “The open set recognition is naturally related to some other problem settings such as out-of-distribution detection [36,32,18], outlier detection [28], and novelty detection [25], etc.”
Patch SVDD: Patch-level SVDD for Anomaly Detection and Segmentation (2020, methodology): “Auto encoder-based methods Early deep learning approaches to anomaly detection used auto encoders [10,11,12].”
A Unified Model for Multi-class Anomaly Detection (2022, background): “5) The feed-forward network is the same as Tab.”
Multiresolution Knowledge Distillation for Anomaly Detection (2020, methodology): “Moreover, anomaly localization is either impossible or poor in most of them [36, 31, 33] and leads to intensive computations that hurt their real-time performance.”
EfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latencies (2023, methodology): “Generative models such as autoencoders [6,8,12,19,31, 41, 52] and GANs [2, 20, 42, 55, 56] have been used extensively for anomaly detection.”
CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances (2020, methodology): “We compare our method with various prior methods including one-class classiﬁer [56, 53], reconstruction-based [55, 48], and self-supervised [15, 26, 1] approaches.”
Anomaly Detection via Reverse Distillation from One-Class Embedding (2022, methodology): “The baselines in this experiment include LSA [1], OCGAN [27], HRN [17], DAAD [16] and MKD [33].”
HRN: A Holistic Approach to One Class Learning (2020, methodology): “Some also use GAN [27] based methods [72, 64, 93, 20].”
Generative-Discriminative Feature Representations for Open-Set Recognition (2020, unknown): “It is also possible to use one-class classification algorithms[24],[21],[20] to solve open-set recognition by modelling known classes.”
SSD: A Unified Framework for Self-Supervised Outlier Detection (2021, background): “While a large number of works focuses on developing methods particularly for single-class modeling in anomaly detection (Perera et al., 2019; Schlegl et al., 2017; Ruff et al., 2018; Chalapathy et al., 2018b; Golan & El-Yaniv, 2018; Wang et al., 2019), some recent work achieve success in both OOD de”
VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization (2021, background): “Reconstruction-based methods are very intuitive and explainable, but their performance is limited when it comes to capture small localized anomalies[21].”