即時語音通信下高效語音感知哈希認證方法研究
[Abstract]:With the rapid development of network communication technology, mobile communication technology and mobile phone communication technology, the convenience of voice acquisition and the maturity of nature and technology have become one of the main ways of instant communication. However, in the process of voice instant communication, speech is vulnerable to the influence of coding and decoding, channel noise, delay and packet loss. Due to the openness of wireless network channel, it opens the door to illegal attacks such as hacking and malicious tampering. In order to achieve efficient speech authentication, and to solve the interaction between robustness, discrimination, authentication efficiency and speech tampering localization, speech perceptual hashing is solved. This paper mainly studies the efficient speech perceptual hashing authentication algorithm for instant speech communication, and implements an efficient speech perceptual hash algorithm with noise reduction, tampering detection and location under the condition of high speed and low reliability of instant communication. The main research contents are as follows: 1. In view of the high efficiency requirement of the perceptual hash algorithm in the instant speech communication, it is necessary for the perceptual hash algorithm to have a small amount of authentication data under the premise of good content keeping operation robustness and different speech discrimination. High operational efficiency. For this reason, an efficient speech perceptual hashing authentication algorithm based on linear predictive minimum mean square error (linear prediction-minimum mean squared) LP-MMSE is proposed. This algorithm not only has good perceptual robustness and different speech distinctiveness to content keeping operation, but also has the advantages of low summarization and high efficiency. In addition, in order to solve the problem of perceptual robustness of content keeping operation, the balance between different speech discrimination and authentication efficiency, an efficient speech perceptual hash authentication algorithm based on improved spectral entropy method is proposed. By comparing the Teager energy operator with the linear prediction coefficient (LP-MMSE) and the linear spectrum method with the (linear spectrum pairn LSP coefficient, the proposed algorithm has good balance in robustness, discrimination and authentication efficiency, and has less perceptual hashing summarization. Can meet the voice communication real-time authentication requirements. 2. In view of the perceptual robustness of voice content retention operation and the distinction of different speech, the requirements of noise reduction, small range tampering and location, and high efficiency authentication are also presented in view of the current voice communication formats of heterogeneous mobile terminals. An efficient speech perceptual hash authentication algorithm based on spectral subtraction and energy entropy ratio is proposed. The algorithm preprocesses the authenticated speech first, then reduces the noise by spectral subtraction; then calculates the energy entropy parameter matrix of each frame by using the energy entropy ratio method after the noise reduction; finally generates binary perceptual hash sequence by hash construction. The experimental results show that the proposed algorithm can effectively reduce the noise of speech, have good perceptual robustness and distinguish different speech, and have less summarization of perceptual hashing and high authentication efficiency. Can achieve small range of tamper detection and location.
【學位授予單位】:蘭州理工大學
【學位級別】:碩士
【學位授予年份】:2017
【分類號】:TN912.3
【參考文獻】
相關期刊論文 前10條
1 隋曉芹;劉培順;;安全高效的移動終端數(shù)字簽名方案[J];計算機系統(tǒng)應用;2016年12期
2 LI Jinfeng;WANG Hongxia;JING Yi;;Audio Perceptual Hashing Based on NMF and MDCT Coefficients[J];Chinese Journal of Electronics;2015年03期
3 李金鳳;吳濤;王宏霞;;基于MFCC相關系數(shù)的語音感知哈希認證算法[J];北京郵電大學學報;2015年02期
4 黃羿博;張秋余;袁占亭;楊仲平;;融合MFCC和LPCC的語音感知哈希算法[J];華中科技大學學報(自然科學版);2015年02期
5 吳吉義;李文娟;黃劍平;章劍林;陳德人;;移動互聯(lián)網(wǎng)研究綜述[J];中國科學:信息科學;2015年01期
6 錢清;王宏霞;劉正輝;;基于基音周期的語音內(nèi)容認證算法[J];鐵道學報;2014年09期
7 楊婉霞;余暉;胡萍;;在壓縮語音編碼中集成信息隱藏方法研究[J];電子學報;2014年07期
8 章麗;張江鑫;;G.729A閉環(huán)基音搜索算法的優(yōu)化與實現(xiàn)[J];杭州電子科技大學學報;2013年05期
9 張金全;王宏霞;李學華;;基于鄰域平均的魯棒音頻水印算法[J];鐵道學報;2012年07期
10 王琳;李成榮;;一種基于自適應譜熵的端點檢測改進方法[J];計算機仿真;2010年12期
相關博士學位論文 前7條
1 游大濤;基于聽覺機理的魯棒特征提取及在說話人識別中的應用[D];哈爾濱工業(yè)大學;2013年
2 王琳琳;說話人識別中的時變魯棒性問題研究[D];清華大學;2013年
3 黃雄華;基于內(nèi)容的數(shù)字音頻水印技術研究[D];西南交通大學;2012年
4 王爾玉;基于若干聲紋信息空間的說話人識別技術研究[D];中國科學技術大學;2012年
5 姜濤;網(wǎng)絡環(huán)境下說話人識別關鍵技術研究[D];哈爾濱工業(yè)大學;2011年
6 焦玉華;音頻感知哈希算法研究[D];哈爾濱工業(yè)大學;2010年
7 古今;語音感知認證的關鍵技術研究[D];中國科學技術大學;2009年
相關碩士學位論文 前10條
1 郝廣洋;語音感知哈希及其在密文語音檢索中的應用研究[D];西南交通大學;2015年
2 張珂;基于G.729A的多路語音處理系統(tǒng)的設計與實現(xiàn)[D];西安科技大學;2014年
3 胡振平;中國聯(lián)通應對即時通信挑戰(zhàn)的策略研究[D];新疆大學;2014年
4 張瑋;基于感知哈希的語音檢索與遠程認證研究[D];西南交通大學;2014年
5 劉揚威;壓縮域語音感知哈希認證算法及安全性研究[D];蘭州理工大學;2014年
6 邸燕君;基于感知哈希的語音內(nèi)容認證方法研究[D];蘭州理工大學;2013年
7 劉大龍;基于感知哈希的音樂聲紋檢索[D];南京郵電大學;2013年
8 李旭飛;說話人識別在身份認證中的應用研究[D];西安電子科技大學;2013年
9 肖星星;短時語音說話人識別方法及應用[D];復旦大學;2012年
10 邱勇;基于感知哈希的語音身份及內(nèi)容認證技術研究[D];西南交通大學;2012年
,本文編號:2186761
本文鏈接:http://sikaile.net/kejilunwen/xinxigongchenglunwen/2186761.html