本文選題：聲源辨別　切入點(diǎn)：聲音信號(hào)分離　出處：《江蘇科技大學(xué)》2017年碩士論文

【摘要】：人們聽到的聲音往往都是由多個(gè)聲音混合而成的,如何從混合的聲音信號(hào)中快速而準(zhǔn)確的分辨出感興趣的聲音信號(hào),一直是研究的熱點(diǎn)。傳統(tǒng)的方法可以進(jìn)行簡(jiǎn)單的聲源辨別,但是當(dāng)涉及到大數(shù)據(jù)量的聲音信號(hào)處理時(shí),影響了其應(yīng)用的實(shí)時(shí)性和準(zhǔn)確性。隨著人工智能時(shí)代的到來(lái),以深度學(xué)習(xí)和GPU并行計(jì)算為代表的新技術(shù)為大數(shù)據(jù)量的聲音信號(hào)處理提供了解決思路,為此本文設(shè)計(jì)了混合聲音信號(hào)辨別的并行化方法,并開展了以下工作:1.分析了國(guó)內(nèi)外混合聲音信號(hào)的研究現(xiàn)狀以及發(fā)展趨勢(shì),以混合聲音信號(hào)為切入點(diǎn),學(xué)習(xí)了混合聲音信號(hào)辨別和GPU并行計(jì)算的相關(guān)知識(shí),并研究了混合聲音信號(hào)分離以及聲源辨別的常用方法。2.對(duì)混合聲音信號(hào)進(jìn)行去均值和白化等預(yù)處理,選取基于負(fù)熵的Fast-ICA算法進(jìn)行混合聲音信號(hào)分離,通過(guò)分析混合聲音信號(hào)分離過(guò)程尋找制約其快速分離的原因,并利用GPU并行化進(jìn)行加速改進(jìn)。3.對(duì)分離后的聲音信號(hào)進(jìn)行多特征值提取,并將提取出的特征值進(jìn)行融合組成復(fù)合特征值,再進(jìn)行聲源辨別。在辨別過(guò)程中,由于傳統(tǒng)神經(jīng)網(wǎng)絡(luò)存在學(xué)習(xí)能力不足的問(wèn)題,針對(duì)這個(gè)缺陷,引入了基于深度信念網(wǎng)絡(luò)(DBN)的聲源辨別模型,以提升混合聲音信號(hào)辨別的準(zhǔn)確率。4.由于要進(jìn)行大數(shù)據(jù)量的聲音信號(hào)處理,并且聲音信號(hào)在處理過(guò)程中同時(shí)又具有方法一致、獨(dú)立性強(qiáng)的特點(diǎn),于是采用GPU并行化方法分別對(duì)基于負(fù)熵的Fast ICA算法、特征值提取和深度信念網(wǎng)絡(luò)模型的訓(xùn)練過(guò)程等操作進(jìn)行優(yōu)化,提高了混合聲音信號(hào)辨別方法的處理效率。通過(guò)仿真和實(shí)驗(yàn)驗(yàn)證,利用GPU并行化對(duì)混合聲音信號(hào)的辨別方法進(jìn)行優(yōu)化改進(jìn),提高了混合聲音信號(hào)分離和辨別的效率,滿足了實(shí)時(shí)性要求。同時(shí),采用基于多特征值融合的復(fù)合特征值作為輸入數(shù)據(jù)和基于深度信念網(wǎng)絡(luò)的聲源辨別模型,提升了混合聲音信號(hào)辨別的準(zhǔn)確率。
[Abstract]:The sound that people hear is often composed of multiple sounds. How to quickly and accurately distinguish the interesting sound signal from the mixed sound signal has always been a hot research topic.The traditional method can distinguish the sound source easily, but it affects the real time and the accuracy of the application when dealing with the sound signal processing of the large amount of data.With the arrival of the era of artificial intelligence, the new technology, represented by deep learning and GPU parallel computing, provides a solution for the sound signal processing of large amount of data. In this paper, a parallelization method for the discrimination of mixed sound signals is designed.And carried out the following work: 1.This paper analyzes the research status and development trend of mixed sound signal at home and abroad. Taking mixed sound signal as the starting point, we study the related knowledge of mixed sound signal discrimination and GPU parallel computing.The common methods of mixed sound signal separation and sound source discrimination. 2.The mixed sound signal is pretreated with de-mean and whitening, and the Fast-ICA algorithm based on negative entropy is selected to separate the mixed sound signal. The reason for the fast separation of mixed sound signal is found by analyzing the separation process of mixed sound signal.And using GPU parallelization to accelerate the improvement. 3.The separated sound signal is extracted with multiple eigenvalues, and the extracted eigenvalues are fused to form composite eigenvalues, and then sound source identification is carried out.Due to the deficiency of learning ability in traditional neural networks, a sound source discrimination model based on deep belief network (DBN) is introduced to improve the accuracy of mixed sound signal discrimination.The extraction of eigenvalues and the training process of the depth belief network model are optimized to improve the processing efficiency of the mixed sound signal identification method.Through simulation and experimental verification, the GPU parallelization is used to optimize and improve the discrimination method of mixed sound signals, which improves the efficiency of separation and discrimination of mixed sound signals and meets the real-time requirements.At the same time, the composite eigenvalue based on multi-eigenvalue fusion is used as input data and the sound source discrimination model based on deep belief network is used to improve the accuracy of mixed sound signal identification.
【學(xué)位授予單位】：江蘇科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TN912.3

【參考文獻(xiàn)】

相關(guān)期刊論文 前8條

1 張春霞;姬楠楠;王冠偉;;受限波爾茲曼機(jī)[J];工程數(shù)學(xué)學(xué)報(bào);2015年02期

2 張鳳儀;夏秀渝;冉國(guó)敬;何禮;葉于林;;多聲源環(huán)境下的魯棒說(shuō)話人識(shí)別[J];計(jì)算機(jī)系統(tǒng)應(yīng)用;2015年04期

3 駱巖紅;萬(wàn)國(guó)峰;王建華;;基于CUDA架構(gòu)的FFT并行計(jì)算研究[J];自動(dòng)化與儀器儀表;2014年12期

4 楊柳;劉鐵英;;CPU浮點(diǎn)運(yùn)算和GPU的通用計(jì)算的技術(shù)差異[J];中國(guó)外資;2013年24期

5 劉雅琴;智愛娟;;幾種語(yǔ)音識(shí)別特征參數(shù)的研究[J];計(jì)算機(jī)技術(shù)與發(fā)展;2009年12期

6 左顥睿;張啟衡;徐勇;趙汝進(jìn);;基于GPU的并行優(yōu)化技術(shù)[J];計(jì)算機(jī)應(yīng)用研究;2009年11期

7 張偉偉;楊鼎才;;用于說(shuō)話人識(shí)別的MFCC的改進(jìn)算法[J];電子測(cè)量技術(shù);2009年08期

8 ;Natural gradient-based recursive least-squares algorithm for adaptive blind source separation[J];Science in China(Series F:Information Sciences);2004年01期

相關(guān)會(huì)議論文 前1條

1 彭詩(shī)雅;;聲紋識(shí)別技術(shù)研究[A];第十六屆全國(guó)青年通信學(xué)術(shù)會(huì)議論文集（上）[C];2011年

相關(guān)碩士學(xué)位論文 前3條

1 呂超;聲源辨別及定位的并行化方法的研究與實(shí)現(xiàn)[D];江蘇科技大學(xué);2016年

2 陸軍建;基于CUDA和深度置信網(wǎng)絡(luò)的手寫字符處理應(yīng)用[D];華東理工大學(xué);2015年

3 王櫻;基于CUDA的FFT并行計(jì)算研究[D];湖南大學(xué);2012年

，

本文編號(hào)：1722077