查詢結果分析
來源資料
相關文獻
- 西文語音辨識系統之設計研究
- 塔米爾文語音辨識系統之設計研究
- 以連續型隱藏式馬可夫模型來計算中文簽名之動態相似度值
- An Overview of RNN-Based Mandarin Speech Recognition Approaches
- 以色彩、紋理、和外形為內容之影像擷取方法
- 基於VQ/HMM之國語文句翻語音中音節音長與振幅參數產生之方法
- An Overview of Mandarin-Speech Tone Recognition
- A Survey on Automatic Speech Recognition with an Illustrative Example on Continuous Speech Recognition of Mandarin
- 中文連續音節辨識之研究
- 隱藏式馬可夫模型介紹
頁籤選單縮合
題 名 | 西文語音辨識系統之設計研究=A Design of Spanish Speech Recognition System |
---|---|
作 者 | 陳志堅; 史世洲; | 書刊名 | 資訊科技國際期刊 |
卷 期 | 6:1 2012.06[民101.06] |
頁 次 | 頁18-25 |
分類號 | 312.8454 |
關鍵詞 | 西文語音辨識; 梅爾頻率倒頻譜係數; 線性預估倒頻譜係數; 隱藏式馬可夫模型; Spanish speech recognition; Linear predictive cepstral coefficients; Mel-frequency cepstral coefficients; Hidden Markov model; |
語 文 | 中文(Chinese) |
中文摘要 | 本論文探討西文語音辨識系統之設計與實作策略。吾人依照西語的發音規則,挑選出 242 個常用單音節之語音特徵,作為主要的訓練與辨識方式。將每個常用單音節,每輪每次以連續唸兩個相同單音的方式錄音,來作為西語語詞的特性依據。吾人錄製六輪共十二次的聲紋特性,來建立西文語音之訓練資料。系統採用梅爾頻率倒頻譜係數及線性預估倒頻譜係數,來萃取單音節語音之特徵參數;運用隱藏式馬可夫模型,作為單音之辨識模型。在 CPU時脈為 1.6GHz的 AMD Sempron 2800+ 之個人電腦與 Ubuntu 9.04作業系統下,針對 4217筆西文語詞,吾人約可達到 86%之正確辨識率,平均所需辨識時間約在 1.5秒以內。而本系統所需的訓練時間約為兩小時。 |
英文摘要 | This paper investigates the design and implementation strategies for a Spanish speech recognition system. It utilizes the speech features of the 242 common Spanish mono-syllables as the major training and recognition methodology. A training database of 12 utterances per mono-syllable is established by applying Spanish pronunciation rules. These 12 utterances are collected through reading 6 rounds of the same mono-syllables twice. Mel-frequency cepstral coefficients, linear predictive cepstral coefficients, and hidden Markov model are used as the two feature models and the recognition model respectively. Under the AMD Sempron Processor 2800+ with 1.6GHz clock rate personal computer and Ubuntu 9.04 operating system environment, a correct phrase recognition rate of 86% can be reached for a 4217 Spanish phrase database. The average computation time for each phrase is about 1.5 seconds, and the training time for the system is about two hours. |
本系統中英文摘要資訊取自各篇刊載內容。