This article handles possibilities of integrating speech technology in to robust wireless technology, allowing voice input for wireless devices. To improve the robustness of speech frontends we introduce, in this paper, a new set of feature vector which is estimated according to the impact of the proposed multidimensional acoustical features on the performance of the Mel-frequency based-advanced frontend. From the denoised acoustic frame using the wiener filter, we optimize the stream weights of multi-in a multi-stream scheme using Karhunen-Loeve Transform (KLT). The proposed frontend is shown to exhibit a stream HMM (Hidden Markov Model) by deploying a discriminative approach based in Likelihood-Ratio Maximization (LRM). Finally, this feature are adequately transformed and reduced relative error rate reduction and provides comparable recognition performance compared with the current DSR-FE (Distributed Speech Recognition FrontEnd) available in wireless communication systems.
{"title":"Acoustic Features for Robust ASR in Cellular Network Applications","authors":"D. Addou, M. Boudraa, B. Boudraa","doi":"10.1109/GSCIT.2016.11","DOIUrl":"https://doi.org/10.1109/GSCIT.2016.11","url":null,"abstract":"This article handles possibilities of integrating speech technology in to robust wireless technology, allowing voice input for wireless devices. To improve the robustness of speech frontends we introduce, in this paper, a new set of feature vector which is estimated according to the impact of the proposed multidimensional acoustical features on the performance of the Mel-frequency based-advanced frontend. From the denoised acoustic frame using the wiener filter, we optimize the stream weights of multi-in a multi-stream scheme using Karhunen-Loeve Transform (KLT). The proposed frontend is shown to exhibit a stream HMM (Hidden Markov Model) by deploying a discriminative approach based in Likelihood-Ratio Maximization (LRM). Finally, this feature are adequately transformed and reduced relative error rate reduction and provides comparable recognition performance compared with the current DSR-FE (Distributed Speech Recognition FrontEnd) available in wireless communication systems.","PeriodicalId":295398,"journal":{"name":"2016 Global Summit on Computer & Information Technology (GSCIT)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130204482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Several studies have proposed to associate a terminological or linguistic part to the ontology's, in order to a clear distinction between the terminological and conceptual component, In particular defined the Ontological and Terminological Resource (OTR). Because domain knowledge can evolve, ontology must be enriched with new knowledge. Methods and tools have been developed in order to extract and organize knowledge explicit and / or implicit, in the text. They generally use specific techniques from different fields of research, especially NLP, machine learning, text mining. However, there is not yet a method / tool that have proven effective for semiautomatic enrichment OTR from text. This paper offers an original solution based on the calculated similarity chain, external semantic resources and ontology's online to enrich OTR from texts. This work is concrétise by a tool called OntoEnrich.
{"title":"Ontological and Terminological Ressource Enrichment from Text Copora","authors":"Anis Tissaoui, Anwer Mezni","doi":"10.1109/GSCIT.2016.18","DOIUrl":"https://doi.org/10.1109/GSCIT.2016.18","url":null,"abstract":"Several studies have proposed to associate a terminological or linguistic part to the ontology's, in order to a clear distinction between the terminological and conceptual component, In particular defined the Ontological and Terminological Resource (OTR). Because domain knowledge can evolve, ontology must be enriched with new knowledge. Methods and tools have been developed in order to extract and organize knowledge explicit and / or implicit, in the text. They generally use specific techniques from different fields of research, especially NLP, machine learning, text mining. However, there is not yet a method / tool that have proven effective for semiautomatic enrichment OTR from text. This paper offers an original solution based on the calculated similarity chain, external semantic resources and ontology's online to enrich OTR from texts. This work is concrétise by a tool called OntoEnrich.","PeriodicalId":295398,"journal":{"name":"2016 Global Summit on Computer & Information Technology (GSCIT)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2016-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124339409","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：