{"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/bagan-data-augmentation-with-balancing-gan","title":"BAGAN: Data Augmentation with Balancing GAN","arxiv_id":"1803.09655","date":"2018-03-26","proceeding":null,"authors":["Giovanni Mariani","Florian Scheidegger","Roxana Istrate","Costas Bekas","Cristiano Malossi"],"abstract":"Image classification datasets are often imbalanced, characteristic that\nnegatively affects the accuracy of deep-learning classifiers. In this work we\npropose balancing GAN (BAGAN) as an augmentation tool to restore balance in\nimbalanced datasets. This is challenging because the few minority-class images\nmay not be enough to train a GAN. We overcome this issue by including during\nthe adversarial training all available images of majority and minority classes.\nThe generative model learns useful features from majority classes and uses\nthese to generate images for minority classes. We apply class conditioning in\nthe latent space to drive the generation process towards a target class. The\ngenerator in the GAN is initialized with the encoder module of an autoencoder\nthat enables us to learn an accurate class-conditioning in the latent space. We\ncompare the proposed methodology with state-of-the-art GANs and demonstrate\nthat BAGAN generates images of superior quality when trained with an imbalanced\ndataset.","url_abs":"http://arxiv.org/abs/1803.09655v2","url_pdf":"http://arxiv.org/pdf/1803.09655v2.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":"bagan-data-augmentation-with-balancing-gan","repo_url":"https://github.com/AhmedImtiazPrio/BAGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"EPL-1.0"}},{"paper_slug":"bagan-data-augmentation-with-balancing-gan","repo_url":"https://github.com/GH920/Improve-Medical-Image-Classification-with-GAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"bagan-data-augmentation-with-balancing-gan","repo_url":"https://github.com/GH920/improved-bagan-gp","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"bagan-data-augmentation-with-balancing-gan","repo_url":"https://github.com/IBM/BAGAN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"EPL-1.0"}}],"tasks":[{"task_slug":"data-augmentation","task_name":"Data Augmentation"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1803.09655","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}