{"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/hpatches-a-benchmark-and-evaluation-of","title":"HPatches: A benchmark and evaluation of handcrafted and learned local descriptors","arxiv_id":"1704.05939","date":"2017-04-19","proceeding":"CVPR 2017 7","authors":["Vassileios Balntas","Karel Lenc","Andrea Vedaldi","Krystian Mikolajczyk"],"abstract":"In this paper, we propose a novel benchmark for evaluating local image\ndescriptors. We demonstrate that the existing datasets and evaluation protocols\ndo not specify unambiguously all aspects of evaluation, leading to ambiguities\nand inconsistencies in results reported in the literature. Furthermore, these\ndatasets are nearly saturated due to the recent improvements in local\ndescriptors obtained by learning them from large annotated datasets. Therefore,\nwe introduce a new large dataset suitable for training and testing modern\ndescriptors, together with strictly defined evaluation protocols in several\ntasks such as matching, retrieval and classification. This allows for more\nrealistic, and thus more reliable comparisons in different application\nscenarios. We evaluate the performance of several state-of-the-art descriptors\nand analyse their properties. We show that a simple normalisation of\ntraditional hand-crafted descriptors can boost their performance to the level\nof deep learning based descriptors within a realistic benchmarks evaluation.","url_abs":"http://arxiv.org/abs/1704.05939v1","url_pdf":"http://arxiv.org/pdf/1704.05939v1.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":[],"tasks":[{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[],"datasets_introduced":[{"slug":"hpatches","name":"HPatches","full_name":"Homography-patches dataset"}],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1704.05939","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}