Metric Learning for Novelty and Anomaly Detection

By Marc Masana, Idoia Ruiz, Joan Serrat, Joost van de Weijer, Antonio M. López. 2018

85 citations recorded by the source corpus.

When neural networks process images which do not resemble the distribution seen during training, so called out-of-distribution images, they often make wrong predictions, and do so too confidently. The capability to detect out-of-distribution images is therefore crucial for many real-world applications. We divide out-of-distribution detection between novelty detection ---images of classes which are not in the training set but are related to those---, and anomaly detection ---images with classes which are unrelated to the training set. By related we mean they contain the same type of objects, like digits in MNIST and SVHN. Most existing work has focused on anomaly detection, and has addressed this problem considering networks trained with the cross-entropy loss. Differently from them, we propose to use metric learning which does not have the drawback of the softmax layer (inherent to cross-entropy methods), which forces the network to divide its prediction power over the learned classes. We perform extensive experiments and evaluate both novelty and anomaly detection, even in a relevant application such as traffic sign recognition, obtaining comparable or better results than previous works.

Subfields: Artificial Intelligence, Radiology, Nuclear Medicine and Imaging

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 Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Semantic Drift Compensation for Class-Incremental Learning (2020, background): “Embedding networks were found to be more robust to the exposure of adversarial examples, and better in the detection of out-of-distribution examples [25, 32].”
Learning Metrics From Teachers: Compact Networks for Image Embedding (2019, methodology): “Embeddings were found efficient on the tasks of out-of-distribution detection [19] and transfer learning [26].”
Characterising Bias in Compressed Models (2020, background): “Leveraging a subset of data points to understand model behaviour or to audit a dataset ﬁts into a broader literature that aims to characterize input data points as prototypes – “most typical" examples of a class – (Carlini et al., 2019; Agarwal & Hooker, 2020; Stock & Cisse, 2017; Jiang et al., 2020”
Recent Advances in Open Set Recognition: A Survey (2018, background): “In fact, as the KUCs’ data is often available at hand [25], [114], [115], we can fully leverage them to reduce such a risk and further improve the robustness of the OSR models for UUCs.”
Student-Teacher Feature Pyramid Matching for Anomaly Detection (2021, background): “Such deep features prove very effective in capturing the intrinsic characteristics of the normal data manifold [3, 10, 24, 32, 47].”
Uninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings (2019, background): “Recently, efforts have been made to improve anomaly detection for one-class or multi-class classification [2, 3, 10, 11, 21, 28, 29].”
Generalized Out-of-Distribution Detection: A Survey (2021, unknown): “2: Distance-based Methods [173], [174], [175], [176], [177], [178], [179]”
Generalized ODIN: Detecting Out-of-Distribution Image Without Learning From Out-of-Distribution Data (2020, methodology): “Some previous works have done a similar attempt to learn without OoD data, such as [35], which uses word embeddings for extra supervision, and [25] which applies metric learning criteria.”
Selective Brain Damage: Measuring the Disparate Impact of Model Pruning (2019, background): “…ﬁts into a broader and non-overlapping body of literature that aims to classify input data points as prototypes – “most typical” examples of a class – ([3, 63]) or outside of the training distribution (OOD) [30, 46, 49, 46, 52] and work on calibrating deep neural network predictions [45, 20, 40].”
Contrastive Training for Improved Out-of-Distribution Detection (2020, unknown): “[22].”