本文選題：語音端點(diǎn)檢測 + 能量聚類�。� 參考：《上海師范大學(xué)》2017年碩士論文

【摘要】：語音端點(diǎn)檢測的目的是檢測出語音信號(hào)中的語音與非語音片段。在很多先進(jìn)的語音處理應(yīng)用的前端處理部分,比如語音識(shí)別,聲紋識(shí)別和語音傳輸,語音端點(diǎn)檢測都是重要的步驟。在所有語音端點(diǎn)檢測系統(tǒng)中,基于能量的語音端點(diǎn)檢測最常被使用�；谀芰康恼Z音端點(diǎn)檢測在無噪聲環(huán)境下性能較好,但是在噪聲環(huán)境下性能下降較多。自適應(yīng)語音端點(diǎn)檢測與傳統(tǒng)的基于能量的語音端點(diǎn)檢測相比,具有很多方面的優(yōu)勢(shì)。然而,自適應(yīng)語音端點(diǎn)檢測中,唯一的最低能量門限不能適應(yīng)不同的噪聲背景。本文的第一個(gè)研究內(nèi)容,提出了一種方法改進(jìn)這個(gè)問題,一種基于k-means的平均能量聚類方法,可以為每個(gè)語音找到更適合的最低能量門限。此外,實(shí)驗(yàn)中還使用了中值濾波,以平滑短時(shí)噪音產(chǎn)生的干擾。在NIST SRE2006說話人測評(píng)(SRE)數(shù)據(jù)上的實(shí)驗(yàn)表明,我們提出的方法比傳統(tǒng)基于能量的VAD和自適應(yīng)VAD均能獲得更好的性能。基于深度神經(jīng)網(wǎng)絡(luò)的語音端點(diǎn)檢測方法由于性能顯著優(yōu)于其他方法,成為近期的研究焦點(diǎn)。本文的第二個(gè)研究內(nèi)容,以一種基于深度神經(jīng)網(wǎng)絡(luò)的語音端點(diǎn)檢測方法為基礎(chǔ),針對(duì)其在低信噪比環(huán)境中表現(xiàn)不佳的問題和易受短時(shí)噪音干擾的問題,分別使用了譜減法語音增強(qiáng)和自適應(yīng)中值濾波的方法做了改進(jìn)。另外,本實(shí)驗(yàn)提出一種監(jiān)督學(xué)習(xí)規(guī)則,類比于人類學(xué)習(xí)先易后難的原則對(duì)神經(jīng)網(wǎng)絡(luò)進(jìn)行訓(xùn)練,顯著加快了神經(jīng)網(wǎng)絡(luò)的收斂速度。在AURORA2數(shù)據(jù)庫上的實(shí)驗(yàn)結(jié)果表明,相比于基線系統(tǒng),改進(jìn)后的方法不僅加速了訓(xùn)練速度,而且還取得了31.12%的相對(duì)性能提升。
[Abstract]:Speech endpoint detection is to detect the speech signal in speech and non speech segments. In many advanced voice processing application front-end processing, such as speech recognition, voice recognition and voice transmission, speech endpoint detection is an important step in all speech endpoint detection system, speech endpoint detection is often the most energy by using the energy based endpoint. Better detection performance in noise environment based on performance, but in the noise environment decreased more. The traditional speech endpoint detection and adaptive speech endpoint detection based on energy ratio, has many advantages. However, adaptive speech endpoint detection, can not only meet the minimum energy threshold noise is different. The first research content, proposes a method to improve this problem, an average energy clustering method based on k-means, You can find the lowest energy threshold is more suitable for each speech. In addition, the median filter is used to smooth short-term interference noise. In NIST SRE2006 (SRE) speaker evaluation indicates that the data on the experiment, we propose a method based on energy performance than the traditional VAD and adaptive VAD can get better. Speech endpoint detection method based on neural network depth due to performance significantly better than other methods, has become the focus of research in recent years. Second the research content of this paper, in a speech endpoint detection method based on deep neural network based on the low SNR environment of poor performance and vulnerable to short-term noise interference the problem, using method of spectral subtraction speech enhancement and adaptive median filtering is improved. In addition, this study proposes a supervised learning rules, analogous to the human learning first The principle of easy and difficult to train the neural network significantly accelerates the convergence speed of neural network. Experimental results on AURORA2 database show that compared with baseline system, the improved method not only speeds up the training speed, but also achieves a relative performance improvement of 31.12%.

【學(xué)位授予單位】：上海師范大學(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4條

1 馮璐;陳威兵;吳宇;;基于語音拖音段的端點(diǎn)檢測算法研究[J];計(jì)算機(jī)工程與科學(xué);2012年10期

2 劉華平;李昕;徐柏齡;姜寧;;語音信號(hào)端點(diǎn)檢測方法綜述及展望[J];計(jì)算機(jī)應(yīng)用研究;2008年08期

3 王書詔;邱天爽;;說話人識(shí)別研究綜述[J];電聲技術(shù);2007年01期

4 王讓定,柴佩琪;一個(gè)基于譜熵的語音端點(diǎn)檢測改進(jìn)方法[J];信息與控制;2004年01期

相關(guān)會(huì)議論文 前1條

1 賈川;張健;陳振標(biāo);徐波;;噪聲環(huán)境下的端點(diǎn)檢測算法研究[A];第六屆全國人機(jī)語音通訊學(xué)術(shù)會(huì)議（NCMMSC6）論文集[C];2001年

相關(guān)碩士學(xué)位論文 前1條

1 周雷;基于聲紋識(shí)別的說話人身份確認(rèn)方法的研究[D];上海師范大學(xué);2016年

，

本文編號(hào)：1742554