{"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/augmentation-for-small-object-detection","title":"Augmentation for small object detection","arxiv_id":"1902.07296","date":"2019-02-19","proceeding":null,"authors":["Mate Kisantal","Zbigniew Wojna","Jakub Murawski","Jacek Naruniec","Kyunghyun Cho"],"abstract":"In recent years, object detection has experienced impressive progress.\nDespite these improvements, there is still a significant gap in the performance\nbetween the detection of small and large objects. We analyze the current\nstate-of-the-art model, Mask-RCNN, on a challenging dataset, MS COCO. We show\nthat the overlap between small ground-truth objects and the predicted anchors\nis much lower than the expected IoU threshold. We conjecture this is due to two\nfactors; (1) only a few images are containing small objects, and (2) small\nobjects do not appear enough even within each image containing them. We thus\npropose to oversample those images with small objects and augment each of those\nimages by copy-pasting small objects many times. It allows us to trade off the\nquality of the detector on large objects with that on small objects. We\nevaluate different pasting augmentation strategies, and ultimately, we achieve\n9.7\\% relative improvement on the instance segmentation and 7.1\\% on the object\ndetection of small objects, compared to the current state of the art method on\nMS COCO.","url_abs":"http://arxiv.org/abs/1902.07296v1","url_pdf":"http://arxiv.org/pdf/1902.07296v1.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":"augmentation-for-small-object-detection","repo_url":"https://github.com/delixing/Data-Augmentation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"augmentation-for-small-object-detection","repo_url":"https://github.com/finepix/small_object_augmentation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"augmentation-for-small-object-detection","repo_url":"https://github.com/siddhanthaldar/PyTorch_Object_Detection","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"augmentation-for-small-object-detection","repo_url":"https://github.com/zzl-pointcloud/Data_Augmentation_Zoo_for_Object_Detection","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"augmentation-for-small-object-detection","repo_url":"https://github.com/2024-MindSpore-1/Code3/tree/main/ResNet18","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"mindspore","reach":null}],"tasks":[{"task_slug":"instance-segmentation","task_name":"Instance Segmentation"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"small-object-detection","task_name":"Small Object Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1902.07296","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}