{"title":"On Homotopy Continuation for Speech Restoration.","authors":"Darian M Onchis, Pedro Real","doi":"10.1007/978-3-319-39441-1_14","DOIUrl":null,"url":null,"abstract":"<p><p>In this paper, a homotopy-based method is employed for the recovery of speech recordings from missing or corrupted samples taken in a noisy environment. The model for the acquisition device is a compressed sensing scenario using Gabor frames. To recover an approximation of the speech file, we used the basis pursuit denoising method with the homotopy continuation algorithm. We tested the proposed method with various speech recordings.</p>","PeriodicalId":91856,"journal":{"name":"Computational topology in image context : 6th International Workshop, CTIC 2016, Marseille, France, June 15-17, 2016, Proceedings. CTIC (Conference) (6th : 2016 : Marseille, France)","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2016-06-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://sci-hub-pdf.com/10.1007/978-3-319-39441-1_14","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computational topology in image context : 6th International Workshop, CTIC 2016, Marseille, France, June 15-17, 2016, Proceedings. CTIC (Conference) (6th : 2016 : Marseille, France)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1007/978-3-319-39441-1_14","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
In this paper, a homotopy-based method is employed for the recovery of speech recordings from missing or corrupted samples taken in a noisy environment. The model for the acquisition device is a compressed sensing scenario using Gabor frames. To recover an approximation of the speech file, we used the basis pursuit denoising method with the homotopy continuation algorithm. We tested the proposed method with various speech recordings.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
