{"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/q-space-novelty-detection-with-variational","title":"q-Space Novelty Detection with Variational Autoencoders","arxiv_id":"1806.02997","date":"2018-06-08","proceeding":null,"authors":["Aleksei Vasilev","Vladimir Golkov","Marc Meissner","Ilona Lipp","Eleonora Sgarlata","Valentina Tomassini","Derek K. Jones","Daniel Cremers"],"abstract":"In machine learning, novelty detection is the task of identifying novel\nunseen data. During training, only samples from the normal class are available.\nTest samples are classified as normal or abnormal by assignment of a novelty\nscore. Here we propose novelty detection methods based on training variational\nautoencoders (VAEs) on normal data. Since abnormal samples are not used during\ntraining, we define novelty metrics based on the (partially complementary)\nassumptions that the VAE is less capable of reconstructing abnormal samples\nwell; that abnormal samples more strongly violate the VAE regularizer; and that\nabnormal samples differ from normal samples not only in input-feature space,\nbut also in the VAE latent space and VAE output. These approaches, combined\nwith various possibilities of using (e.g. sampling) the probabilistic VAE to\nobtain scalar novelty scores, yield a large family of methods. We apply these\nmethods to magnetic resonance imaging, namely to the detection of\ndiffusion-space (q-space) abnormalities in diffusion MRI scans of multiple\nsclerosis patients, i.e. to detect multiple sclerosis lesions without using any\nlesion labels for training. Many of our methods outperform previously proposed\nq-space novelty detection methods. We also evaluate the proposed methods on the\nMNIST handwritten digits dataset and show that many of them are able to\noutperform the state of the art.","url_abs":"http://arxiv.org/abs/1806.02997v2","url_pdf":"http://arxiv.org/pdf/1806.02997v2.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":[{"paper_slug":"q-space-novelty-detection-with-variational","repo_url":"https://github.com/VAlex22/ND_VAE","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"diffusion-mri","task_name":"Diffusion  MRI"},{"task_slug":"novelty-detection","task_name":"Novelty Detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}