{"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/class-balanced-distillation-for-long-tailed","title":"Class-Balanced Distillation for Long-Tailed Visual Recognition","arxiv_id":"2104.05279","date":"2021-04-12","proceeding":null,"authors":["Ahmet Iscen","André Araujo","Boqing Gong","Cordelia Schmid"],"abstract":"Real-world imagery is often characterized by a significant imbalance of the number of images per class, leading to long-tailed distributions. An effective and simple approach to long-tailed visual recognition is to learn feature representations and a classifier separately, with instance and class-balanced sampling, respectively. In this work, we introduce a new framework, by making the key observation that a feature representation learned with instance sampling is far from optimal in a long-tailed setting. Our main contribution is a new training method, referred to as Class-Balanced Distillation (CBD), that leverages knowledge distillation to enhance feature representations. CBD allows the feature representation to evolve in the second training stage, guided by the teacher learned in the first stage. The second stage uses class-balanced sampling, in order to focus on under-represented classes. This framework can naturally accommodate the usage of multiple teachers, unlocking the information from an ensemble of models to enhance recognition capabilities. Our experiments show that the proposed technique consistently outperforms the state of the art on long-tailed recognition benchmarks such as ImageNet-LT, iNaturalist17 and iNaturalist18.","url_abs":"https://arxiv.org/abs/2104.05279v2","url_pdf":"https://arxiv.org/pdf/2104.05279v2.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":"class-balanced-distillation-for-long-tailed","repo_url":"https://github.com/google-research/google-research","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null},{"paper_slug":"class-balanced-distillation-for-long-tailed","repo_url":"https://github.com/google-research/google-research/tree/master/class_balanced_distillation","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"jax","reach":null},{"paper_slug":"class-balanced-distillation-for-long-tailed","repo_url":"https://github.com/rahulvigneswaran/Class-Balanced-Distillation-for-Long-Tailed-Visual-Recognition.pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"},{"task_slug":"long-tail-learning","task_name":"Long-tail Learning"}],"methods":[{"method_slug":"knowledge-distillation","method_name":"Knowledge Distillation"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-classification-on-inaturalist-2018","task":"Image Classification","dataset":"iNaturalist 2018","model":"CBD-ENS (ResNet-101)","rank_in_archive_order":26,"of":60,"metrics":{"Top-1 Accuracy":"75.3%"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-inaturalist-2018","task":"Image Classification","dataset":"iNaturalist 2018","model":"CBD-ENS (ResNet-50)","rank_in_archive_order":31,"of":60,"metrics":{"Top-1 Accuracy":"73.6%"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-imagenet-lt","task":"Long-tail Learning","dataset":"ImageNet-LT","model":"CBD-ENS (ResNet-152)","rank_in_archive_order":26,"of":69,"metrics":{"Top-1 Accuracy":"57.7"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-imagenet-lt","task":"Long-tail Learning","dataset":"ImageNet-LT","model":"CBD-ENS (ResNet-50)","rank_in_archive_order":36,"of":69,"metrics":{"Top-1 Accuracy":"55.6"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-inaturalist-2018","task":"Long-tail Learning","dataset":"iNaturalist 2018","model":"CBD-ENS (ResNet-101)","rank_in_archive_order":13,"of":43,"metrics":{"Top-1 Accuracy":"75.3%"},"uses_additional_data":false},{"leaderboard":"/sota/long-tail-learning-on-inaturalist-2018","task":"Long-tail Learning","dataset":"iNaturalist 2018","model":"CBD-ENS (ResNet-50)","rank_in_archive_order":21,"of":43,"metrics":{"Top-1 Accuracy":"73.6%"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2104.05279","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}