{"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/bs-nets-an-end-to-end-framework-for-band","title":"BS-Nets: An End-to-End Framework For Band Selection of Hyperspectral Image","arxiv_id":"1904.08269","date":"2019-04-17","proceeding":null,"authors":["Yaoming Cai","Xiaobo Liu","Zhihua Cai"],"abstract":"Hyperspectral image (HSI) consists of hundreds of continuous narrow bands\nwith high spectral correlation, which would lead to the so-called Hughes\nphenomenon and the high computational cost in processing. Band selection has\nbeen proven effective in avoiding such problems by removing the redundant\nbands. However, many of existing band selection methods separately estimate the\nsignificance for every single band and cannot fully consider the nonlinear and\nglobal interaction between spectral bands. In this paper, by assuming that a\ncomplete HSI can be reconstructed from its few informative bands, we propose a\ngeneral band selection framework, Band Selection Network (termed as BS-Net).\nThe framework consists of a band attention module (BAM), which aims to\nexplicitly model the nonlinear inter-dependencies between spectral bands, and a\nreconstruction network (RecNet), which is used to restore the original HSI cube\nfrom the learned informative bands, resulting in a flexible architecture. The\nresulting framework is end-to-end trainable, making it easier to train from\nscratch and to combine with existing networks. We implement two BS-Nets\nrespectively using fully connected networks (BS-Net-FC) and convolutional\nneural networks (BS-Net-Conv), and compare the results with many existing band\nselection approaches for three real hyperspectral images, demonstrating that\nthe proposed BS-Nets can accurately select informative band subset with less\nredundancy and achieve significantly better classification performance with an\nacceptable time cost.","url_abs":"http://arxiv.org/abs/1904.08269v1","url_pdf":"http://arxiv.org/pdf/1904.08269v1.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":"bs-nets-an-end-to-end-framework-for-band","repo_url":"https://github.com/ucalyptus/BS-Nets-Implementation-Pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"bs-nets-an-end-to-end-framework-for-band","repo_url":"https://github.com/AngryCai/BS-Nets","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"hyperspectral-image-classification","task_name":"Hyperspectral Image Classification"}],"methods":[],"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}