{"title":"时域内独立分量的模糊聚类盲音频源分离方法","authors":"J. Málek, Zbyněk Koldovský","doi":"10.1109/IWECMS.2011.5952370","DOIUrl":null,"url":null,"abstract":"This paper deals with several modifications of an existing Blind Audio Source Separation (BASS) method called T-ABCD. The method applies Independent Component Analysis (ICA) in the time-domain, which gives independent components of individual signals that form unknown groups. The need is to recover these groups using a clustering algorithm and a similarity measure, and reconstruct the separated signals from the groups then. In this paper, several novel criteria that are suitable to measure the similarity between audio components are proposed. Next, fuzzy clustering algorithms are applied to group the components, and novel reconstruction approaches relying on proper weighting of components are proposed. The proposed modifications are compared by experiments, and conclusions are drawn.","PeriodicalId":211450,"journal":{"name":"2011 10th International Workshop on Electronics, Control, Measurement and Signals","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2011-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Fuzzy clustering of independent components within time-domain blind audio source separation method\",\"authors\":\"J. Málek, Zbyněk Koldovský\",\"doi\":\"10.1109/IWECMS.2011.5952370\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"This paper deals with several modifications of an existing Blind Audio Source Separation (BASS) method called T-ABCD. The method applies Independent Component Analysis (ICA) in the time-domain, which gives independent components of individual signals that form unknown groups. The need is to recover these groups using a clustering algorithm and a similarity measure, and reconstruct the separated signals from the groups then. In this paper, several novel criteria that are suitable to measure the similarity between audio components are proposed. Next, fuzzy clustering algorithms are applied to group the components, and novel reconstruction approaches relying on proper weighting of components are proposed. The proposed modifications are compared by experiments, and conclusions are drawn.\",\"PeriodicalId\":211450,\"journal\":{\"name\":\"2011 10th International Workshop on Electronics, Control, Measurement and Signals\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2011-06-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2011 10th International Workshop on Electronics, Control, Measurement and Signals\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/IWECMS.2011.5952370\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2011 10th International Workshop on Electronics, Control, Measurement and Signals","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IWECMS.2011.5952370","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Fuzzy clustering of independent components within time-domain blind audio source separation method
This paper deals with several modifications of an existing Blind Audio Source Separation (BASS) method called T-ABCD. The method applies Independent Component Analysis (ICA) in the time-domain, which gives independent components of individual signals that form unknown groups. The need is to recover these groups using a clustering algorithm and a similarity measure, and reconstruct the separated signals from the groups then. In this paper, several novel criteria that are suitable to measure the similarity between audio components are proposed. Next, fuzzy clustering algorithms are applied to group the components, and novel reconstruction approaches relying on proper weighting of components are proposed. The proposed modifications are compared by experiments, and conclusions are drawn.