{"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/latch-learned-arrangements-of-three-patch","title":"LATCH: Learned Arrangements of Three Patch Codes","arxiv_id":"1501.03719","date":"2015-01-15","proceeding":null,"authors":["Gil Levi","Tal Hassner"],"abstract":"We present a novel means of describing local image appearances using binary\nstrings. Binary descriptors have drawn increasing interest in recent years due\nto their speed and low memory footprint. A known shortcoming of these\nrepresentations is their inferior performance compared to larger, histogram\nbased descriptors such as the SIFT. Our goal is to close this performance gap\nwhile maintaining the benefits attributed to binary representations. To this\nend we propose the Learned Arrangements of Three Patch Codes descriptors, or\nLATCH. Our key observation is that existing binary descriptors are at an\nincreased risk from noise and local appearance variations. This, as they\ncompare the values of pixel pairs; changes to either of the pixels can easily\nlead to changes in descriptor values, hence damaging its performance. In order\nto provide more robustness, we instead propose a novel means of comparing pixel\npatches. This ostensibly small change, requires a substantial redesign of the\ndescriptors themselves and how they are produced. Our resulting LATCH\nrepresentation is rigorously compared to state-of-the-art binary descriptors\nand shown to provide far better performance for similar computation and space\nrequirements.","url_abs":"http://arxiv.org/abs/1501.03719v1","url_pdf":"http://arxiv.org/pdf/1501.03719v1.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":"latch-learned-arrangements-of-three-patch","repo_url":"https://github.com/komrad36/CLATCH","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"latch-learned-arrangements-of-three-patch","repo_url":"https://github.com/komrad36/LATCH","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"latch-learned-arrangements-of-three-patch","repo_url":"https://github.com/komrad36/UCLATCH","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"latch-learned-arrangements-of-three-patch","repo_url":"https://github.com/komrad36/ULATCH","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}