{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/ocgan-one-class-novelty-detection-using-gans","title":"OCGAN: One-class Novelty Detection Using GANs with Constrained Latent Representations","arxiv_id":"1903.08550","date":"2019-03-20","proceeding":"CVPR 2019 6","authors":["Pramuditha Perera","Ramesh Nallapati","Bing Xiang"],"abstract":"We present a novel model called OCGAN for the classical problem of one-class\nnovelty detection, where, given a set of examples from a particular class, the\ngoal is to determine if a query example is from the same class. Our solution is\nbased on learning latent representations of in-class examples using a denoising\nauto-encoder network. The key contribution of our work is our proposal to\nexplicitly constrain the latent space to exclusively represent the given class.\nIn order to accomplish this goal, firstly, we force the latent space to have\nbounded support by introducing a tanh activation in the encoder's output layer.\nSecondly, using a discriminator in the latent space that is trained\nadversarially, we ensure that encoded representations of in-class examples\nresemble uniform random samples drawn from the same bounded space. Thirdly,\nusing a second adversarial discriminator in the input space, we ensure all\nrandomly drawn latent samples generate examples that look real. Finally, we\nintroduce a gradient-descent based sampling technique that explores points in\nthe latent space that generate potential out-of-class examples, which are fed\nback to the network to further train it to generate in-class examples from\nthose points. The effectiveness of the proposed method is measured across four\npublicly available datasets using two one-class novelty detection protocols\nwhere we achieve state-of-the-art results.","url_abs":"http://arxiv.org/abs/1903.08550v1","url_pdf":"http://arxiv.org/pdf/1903.08550v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"anomaly-detection","task_name":"Anomaly Detection"},{"task_slug":"denoising","task_name":"Denoising"},{"task_slug":"novelty-detection","task_name":"Novelty Detection"}],"methods":[{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/anomaly-detection-on-hyper-kvasir-dataset","task":"Anomaly Detection","dataset":"Hyper-Kvasir Dataset","model":"OCGAN","rank_in_archive_order":6,"of":6,"metrics":{"AUC":"0.813"},"uses_additional_data":false},{"leaderboard":"/sota/anomaly-detection-on-one-class-cifar-10","task":"Anomaly Detection","dataset":"One-class CIFAR-10","model":"OCGAN","rank_in_archive_order":31,"of":36,"metrics":{"AUROC":"66.83"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1903.08550","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}