{"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/towards-optimal-discrete-online-hashing-with","title":"Towards Optimal Discrete Online Hashing with Balanced Similarity","arxiv_id":"1901.10185","date":"2019-01-29","proceeding":null,"authors":["Mingbao Lin","Rongrong Ji","Hong Liu","Xiaoshuai Sun","Yongjian Wu","Yunsheng Wu"],"abstract":"When facing large-scale image datasets, online hashing serves as a promising\nsolution for online retrieval and prediction tasks. It encodes the online\nstreaming data into compact binary codes, and simultaneously updates the hash\nfunctions to renew codes of the existing dataset. To this end, the existing\nmethods update hash functions solely based on the new data batch, without\ninvestigating the correlation between such new data and the existing dataset.\nIn addition, existing works update the hash functions using a relaxation\nprocess in its corresponding approximated continuous space. And it remains as\nan open problem to directly apply discrete optimizations in online hashing. In\nthis paper, we propose a novel supervised online hashing method, termed\nBalanced Similarity for Online Discrete Hashing (BSODH), to solve the above\nproblems in a unified framework. BSODH employs a well-designed hashing\nalgorithm to preserve the similarity between the streaming data and the\nexisting dataset via an asymmetric graph regularization. We further identify\nthe \"data-imbalance\" problem brought by the constructed asymmetric graph, which\nrestricts the application of discrete optimization in our problem. Therefore, a\nnovel balanced similarity is further proposed, which uses two equilibrium\nfactors to balance the similar and dissimilar weights and eventually enables\nthe usage of discrete optimizations. Extensive experiments conducted on three\nwidely-used benchmarks demonstrate the advantages of the proposed method over\nthe state-of-the-art methods.","url_abs":"http://arxiv.org/abs/1901.10185v2","url_pdf":"http://arxiv.org/pdf/1901.10185v2.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":"towards-optimal-discrete-online-hashing-with","repo_url":"https://github.com/lmbxmu/mycode","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1901.10185","atlas_url":"https://app.syntology.ai/?focus=1901.10185","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}