{"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/punch-positive-unlabelled-classification","title":"PUNCH: Positive UNlabelled Classification based information retrieval in Hyperspectral images","arxiv_id":"1904.04547","date":"2019-04-09","proceeding":"Submitted to ACMMM-2019 2019 4","authors":["Anirban Santara","Jayeeta Datta","Sourav Sarkar","Ankur Garg","Kirti Padia","Pabitra Mitra"],"abstract":"Hyperspectral images of land-cover captured by airborne or satellite-mounted\nsensors provide a rich source of information about the chemical composition of\nthe materials present in a given place. This makes hyperspectral imaging an\nimportant tool for earth sciences, land-cover studies, and military and\nstrategic applications. However, the scarcity of labeled training examples and\nspatial variability of spectral signature are two of the biggest challenges\nfaced by hyperspectral image classification. In order to address these issues,\nwe aim to develop a framework for material-agnostic information retrieval in\nhyperspectral images based on Positive-Unlabelled (PU) classification. Given a\nhyperspectral scene, the user labels some positive samples of a material he/she\nis looking for and our goal is to retrieve all the remaining instances of the\nquery material in the scene. Additionally, we require the system to work\nequally well for any material in any scene without the user having to disclose\nthe identity of the query material. This material-agnostic nature of the\nframework provides it with superior generalization abilities. We explore two\nalternative approaches to solve the hyperspectral image classification problem\nwithin this framework. The first approach is an adaptation of non-negative risk\nestimation based PU learning for hyperspectral data. The second approach is\nbased on one-versus-all positive-negative classification where the negative\nclass is approximately sampled using a novel spectral-spatial retrieval model.\nWe propose two annotator models - uniform and blob - that represent the\nlabelling patterns of a human annotator. We compare the performances of the\nproposed algorithms for each annotator model on three benchmark hyperspectral\nimage datasets - Indian Pines, Pavia University and Salinas.","url_abs":"http://arxiv.org/abs/1904.04547v1","url_pdf":"http://arxiv.org/pdf/1904.04547v1.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":"punch-positive-unlabelled-classification","repo_url":"https://github.com/HSISeg/HSISeg","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"hyperspectral-image-classification","task_name":"Hyperspectral Image Classification"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"information-retrieval","task_name":"Information Retrieval"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":"image-classification","task_name":"image-classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}