{"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/learning-to-count-objects-in-natural-images","title":"Learning to Count Objects in Natural Images for Visual Question Answering","arxiv_id":"1802.05766","date":"2018-02-15","proceeding":"ICLR 2018 1","authors":["Yan Zhang","Jonathon Hare","Adam Prügel-Bennett"],"abstract":"Visual Question Answering (VQA) models have struggled with counting objects\nin natural images so far. We identify a fundamental problem due to soft\nattention in these models as a cause. To circumvent this problem, we propose a\nneural network component that allows robust counting from object proposals.\nExperiments on a toy task show the effectiveness of this component and we\nobtain state-of-the-art accuracy on the number category of the VQA v2 dataset\nwithout negatively affecting other categories, even outperforming ensemble\nmodels with our single model. On a difficult balanced pair metric, the\ncomponent gives a substantial improvement in counting over a strong baseline by\n6.6%.","url_abs":"http://arxiv.org/abs/1802.05766v1","url_pdf":"http://arxiv.org/pdf/1802.05766v1.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":"learning-to-count-objects-in-natural-images","repo_url":"https://github.com/Cyanogenoid/vqa-counting","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"visual-question-answering-1","task_name":"Visual Question Answering"},{"task_slug":"visual-question-answering","task_name":"Visual Question Answering (VQA)"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/visual-question-answering-on-vqa-v2-test-dev","task":"Visual Question Answering (VQA)","dataset":"VQA v2 test-dev","model":"DMN","rank_in_archive_order":35,"of":56,"metrics":{"Accuracy":"68.09"},"uses_additional_data":false},{"leaderboard":"/sota/visual-question-answering-on-vqa-v2-test-std","task":"Visual Question Answering (VQA)","dataset":"VQA v2 test-std","model":"DMN","rank_in_archive_order":32,"of":38,"metrics":{"overall":"68.4"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1802.05766","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}