國立高雄大學圖資館 |

語系: 繁體中文

說明(常見問題)

圖資館首頁

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

基於凸型封閉曲面邊界條件分析之盲訊號分離演算法開發 = Developm...

國立高雄大學電機工程學系碩士班

基於凸型封閉曲面邊界條件分析之盲訊號分離演算法開發 = Development of Blind-Source-Separation Algorithm based on Bounded-Component Analysis Associated with Convex-Hull Approach

紀錄類型:	書目-語言資料,印刷品 : 單行本
並列題名:	Development of Blind-Source-Separation Algorithm based on Bounded-Component Analysis Associated with Convex-Hull Approach
作者:	張惟順,
其他團體作者:	國立高雄大學
出版地:	[高雄市]
出版者:	國立高雄大學;
出版年:	2015[民104]
面頁冊數:	[10]，91葉圖，表 : 30公分;
標題:	盲訊號
標題:	blind signal
電子資源:	https://hdl.handle.net/11296/3nvh55
附註:	109年11月18日公開
摘要註:	本論文目的是用邊界條件分析演算法進行多種類型聲音訊號的盲訊號分離。演算法的部分,我們以Cruce提出處理聲音領域的邊界條件分析演算法為基礎,開發出使用平均濾波器先行處理預白化單元的輸入訊號,並以比較先前找到的兩個頂點與新加入的點形成的面積之方式逐一決定所構成凸型封閉曲面的頂點。模擬時,我們用九種類型的聲音訊號進行分離,原本的邊界條件分析演算法與本論文提出的演算法都能分離Speech4、Speech8和Speech10和25speakerNOSYNC等聲音訊號,其效能參數依序為0.041151、0.023091、0.025838和0.07348;用本論文提出的演算法分離前述訊號得到的效能參數依序為0.043626、0.019537、0.026982和0.075524。此外,本論文提出的演算法也能分離Nband5、ACSin10d和Speech4_sin0_05等三種訊號。本論文提出的演算法透過犧牲部份效能參數,使演算法具有分離更多種類訊號的能力。 In this thesis, the bounded component analysis algorithms were applied to blind signal processing for separating many kinds of signals. For the algorithm, we placed an average filter before the input signals of the prewhitening stage based on bounded component analysis algorithm proposed by Cruces. For the bounded component analysis stage, we iteratively compared each area formed by previous two vertices with a new point to determine whether the new point belongs to the vertices of the convex hull that will be found. We used the nine different source signals for simulation. The bounded component analysis algorithm and this algorithm could separate Speech4, Speech8, Speech10, and 25speakerNOSYNC signals. Bounded component analysis algorithm separated the signals as mentioned with the performance indices of 0.041151, 0.023091, 0.025838, and 0.07348, respectively. The proposed algorithm separated the signals as mentioned with the performance indices of 0.043626, 0.019537, 0.026982, and 0.075524, respectively. In additional to the mentioned four kinds of signals, this algorithm could separate three more kinds of signals, they are Nband5, ACSin10d and Speech4_sin0_05.

基於凸型封閉曲面邊界條件分析之盲訊號分離演算法開發 = Development of Blind-Source-Separation Algorithm based on Bounded-Component Analysis Associated with Convex-Hull Approach
張, 惟順

基於凸型封閉曲面邊界條件分析之盲訊號分離演算法開發 = Development of Blind-Source-Separation Algorithm based on Bounded-Component Analysis Associated with Convex-Hull Approach / 張惟順撰 - [高雄市] : 國立高雄大學, 2015[民104]. - [10]，91葉 ; 圖，表 ; 30公分.
109年11月18日公開.
參考書目：葉68-71.
盲訊號blind signal

基於凸型封閉曲面邊界條件分析之盲訊號分離演算法開發 = Development of Blind-Source-Separation Algorithm based on Bounded-Component Analysis Associated with Convex-Hull Approach
LDR:03429nam0a2200277 450 001 458462
005 20220310133914.0
010 0 $b 精裝
010 0 $b 平裝
100 $a 20151007d2015 k y0chiy50 e
101 0 $a chi $d chi $d eng
102 $a tw
105 $a ak am 000yy
200 1 $a 基於凸型封閉曲面邊界條件分析之盲訊號分離演算法開發 $d Development of Blind-Source-Separation Algorithm based on Bounded-Component Analysis Associated with Convex-Hull Approach $z eng $f 張惟順撰
210 $a [高雄市] $c 國立高雄大學 $d 2015[民104]
215 0 $a [10]，91葉 $c 圖，表 $d 30公分
300 $a 109年11月18日公開
314 $a 指導教授：陳春僥博士
320 $a 參考書目：葉68-71
328 $a 碩士論文--國立高雄大學電機工程學系碩士班
330 $a 本論文目的是用邊界條件分析演算法進行多種類型聲音訊號的盲訊號分離。演算法的部分,我們以Cruce提出處理聲音領域的邊界條件分析演算法為基礎,開發出使用平均濾波器先行處理預白化單元的輸入訊號,並以比較先前找到的兩個頂點與新加入的點形成的面積之方式逐一決定所構成凸型封閉曲面的頂點。模擬時,我們用九種類型的聲音訊號進行分離,原本的邊界條件分析演算法與本論文提出的演算法都能分離Speech4、Speech8和Speech10和25speakerNOSYNC等聲音訊號,其效能參數依序為0.041151、0.023091、0.025838和0.07348;用本論文提出的演算法分離前述訊號得到的效能參數依序為0.043626、0.019537、0.026982和0.075524。此外,本論文提出的演算法也能分離Nband5、ACSin10d和Speech4_sin0_05等三種訊號。本論文提出的演算法透過犧牲部份效能參數,使演算法具有分離更多種類訊號的能力。 In this thesis, the bounded component analysis algorithms were applied to blind signal processing for separating many kinds of signals. For the algorithm, we placed an average filter before the input signals of the prewhitening stage based on bounded component analysis algorithm proposed by Cruces. For the bounded component analysis stage, we iteratively compared each area formed by previous two vertices with a new point to determine whether the new point belongs to the vertices of the convex hull that will be found. We used the nine different source signals for simulation. The bounded component analysis algorithm and this algorithm could separate Speech4, Speech8, Speech10, and 25speakerNOSYNC signals. Bounded component analysis algorithm separated the signals as mentioned with the performance indices of 0.041151, 0.023091, 0.025838, and 0.07348, respectively. The proposed algorithm separated the signals as mentioned with the performance indices of 0.043626, 0.019537, 0.026982, and 0.075524, respectively. In additional to the mentioned four kinds of signals, this algorithm could separate three more kinds of signals, they are Nband5, ACSin10d and Speech4_sin0_05.
510 1 $a Development of Blind-Source-Separation Algorithm based on Bounded-Component Analysis Associated with Convex-Hull Approach $z eng
610 # 0 $a 盲訊號 $a 邊界條件分析 $a 凸型封閉曲面 $a 訊號分離 $a 聲音
610 # 1 $a blind signal $a bounded component analysis $a convex hull $a source
681 $a 008M/0019 $b 542201 1192 $v 2007年版
700 1 $a 張 $b 惟順 $4 撰 $3 709700
712 0 2 $a 國立高雄大學 $b 電機工程學系碩士班 $3 166118
801 0 $a tw $b NUK $c 20201112 $g CCR
856 7 # $u https://hdl.handle.net/11296/3nvh55 $z 電子資源 $2 http