{"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/cut-paste-and-learn-surprisingly-easy","title":"Cut, Paste and Learn: Surprisingly Easy Synthesis for Instance Detection","arxiv_id":"1708.01642","date":"2017-08-04","proceeding":"ICCV 2017 10","authors":["Debidatta Dwibedi","Ishan Misra","Martial Hebert"],"abstract":"A major impediment in rapidly deploying object detection models for instance\ndetection is the lack of large annotated datasets. For example, finding a large\nlabeled dataset containing instances in a particular kitchen is unlikely. Each\nnew environment with new instances requires expensive data collection and\nannotation. In this paper, we propose a simple approach to generate large\nannotated instance datasets with minimal effort. Our key insight is that\nensuring only patch-level realism provides enough training signal for current\nobject detector models. We automatically `cut' object instances and `paste'\nthem on random backgrounds. A naive way to do this results in pixel artifacts\nwhich result in poor performance for trained models. We show how to make\ndetectors ignore these artifacts during training and generate data that gives\ncompetitive performance on real data. Our method outperforms existing synthesis\napproaches and when combined with real images improves relative performance by\nmore than 21% on benchmark datasets. In a cross-domain setting, our synthetic\ndata combined with just 10% real data outperforms models trained on all real\ndata.","url_abs":"http://arxiv.org/abs/1708.01642v1","url_pdf":"http://arxiv.org/pdf/1708.01642v1.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":"cut-paste-and-learn-surprisingly-easy","repo_url":"https://github.com/IliasMAOUDJ/dataset_gen","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"cut-paste-and-learn-surprisingly-easy","repo_url":"https://github.com/a-nau/image-selection-and-cnn-training","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"cut-paste-and-learn-surprisingly-easy","repo_url":"https://github.com/a-nau/synthetic-dataset-generation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"cut-paste-and-learn-surprisingly-easy","repo_url":"https://github.com/animikhaich/single-object-background-subtractor","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}},{"paper_slug":"cut-paste-and-learn-surprisingly-easy","repo_url":"https://github.com/debidatta/syndata-generation","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}},{"paper_slug":"cut-paste-and-learn-surprisingly-easy","repo_url":"https://github.com/gyhandy/text2image-for-detection","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1708.01642","atlas_url":"https://app.syntology.ai/?focus=1708.01642","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}