臺灣博碩士論文加值系統

語音識別的研究日新月異，已普遍應用在不同的領域，不論是手機的語音輸入或是語音助理等。在聲音訊號相關研究中，對於嬰幼兒哭聲之研究也日益成熟，不論是透過哭聲來判斷嬰幼兒相關行為或是辨識哭聲來判斷嬰幼兒的情緒和需求等，相關研究中已有準確率相當高的嬰語翻譯機，但就如其他的語音識別系統一樣，對於多聲源接收依然是一大問題，多數語音識別功能無法處理同時間的多個聲源，為了降低多聲源所造成的高誤判率，本論文提出利用獨立成分分析法來針對多個混合訊號進行分離及還原訊號，實驗結果顯示，透過本系統分離的訊號，對於嬰語翻譯機的誤判率有顯著的下降，大大提升了嬰語翻譯機的辨識率以及擁有辨識多聲源的能力，未使用本系統前之辨識率為34%，使用本系統後之辨識率為68%。

The research on analysing infant crying has received many attentions in recent years. In our prior work, a baby crying translation method called infant crying translator was proposed and showed a high recognition accuracy. However, in a real environment, there may be more than one baby crying. These mixed cries will seriously affect the accuracy of recognition. In order to isolate these mixed cries, the independent component analysis was adopted herein. Experimental results show that the proposed method can separate out the mixed cries and greatly improves the recognition rate of infant crying translator. The recognition rate increased from 34% to 68%.

目錄
摘要 i
ABSTRACT ii
目錄 iii
表目錄 vi
圖目錄 vii
第一章 緒論 1
1.1 研究動機與目的 1
1.2 文獻探討 2
1.3 研究方法 3
1.4 章節大鋼 3
第二章 相關理論 4
2.1 聲音的產生 4
2.1.1 音強(Intensity) 5
2.1.2 音調(Pitch) 6
2.1.3 音色(Timbre) 7
2.2 盲訊號分離(Blind Sources Separation) 8
2.2.1雞尾酒會問題(Cocktail Party Problem) 9
2.3 主成分分析法(Principal Components Analysis） 10
2.4 快速獨立成分分析法(Fast Independent Components Analysis ) 11
2.5 歐幾里得距離(Euclidean Distance) 12
2.6 中央極限定理(Central Limit Theorem) 12
2.7快速傅立葉轉換(Fast Fourier Transform,FFT) 13
2.8 動態時間校正(Dynamic Time Warping) 16
2.9 結構相似性(Structural Similarity) 17
2.10 頻譜(Frequency Spectrum) 17
2.11 嬰幼兒哭聲辨識 18
第三章 研究方法 19
3.1 系統架構 19
3.2 獨立成分分析之資料限制 20
3.3 感測器數量限制 20
3.4 前處理 21
3.4.1 訊號正規化 21
3.4.2 端點偵測 22
3.4.3 中心化(Centering) 22
3.4.4 白化(Whitening) 22
第四章 實驗結果與討論 24
4.1 實驗環境與設計 24
4.2 實驗資料 26
4.3 評估準則 27
4.3.1 訊號時域相似性 28
4.3.2 訊號時域動態時間校正相似性 29
4.3.3 訊號頻域相似性 31
4.3.4 聲譜結構相似性 33
4.4 使用主成分分析 35
4.5 使用快速獨立成分分析分離 41
4.6 分離訊號與還原訊號方法比較 47
4.7 哭聲辨識結果 49
4.8 哭聲辨識結果比較 50
第五章 結論 52
參考文獻 53

[1]J. F. Cardoso, “Blind signal separation: statistical principles,” Proceedings of the IEEE, Vol. 86, No. 10, pp. 2009-2025, 1998.
[2]M. A. Bee, C. Micheyl, “The cocktail party problem: what is it? How can it be solved? And why should animal behaviorists study it?,” Journal of Comparative Psychology, Vol. 122, No. 3, pp. 235, 2008.
[3]J. J. Rieta, F. Castells, C. Sánchez, V. Zarzoso, J. Millet, ”Atrial activity extraction for atrial fibrillation analysis using blind source separation,” IEEE Transactions on Biomedical Engineering, Vol. 51, No. 7, pp. 1176-1186, 2004.
[4]A. Hyvarinen, “Fast and robust fixed-point algorithms for independent component analysis,” IEEE Transactions on Neural Networks, Vol. 10, No. 3, pp. 626-634, 1999.
[5]C. M. Gallippi, K. R. Nightingale, G. E. Trahey, “Blind source separation-based adaptive filtering of physiological and ARFI-induced tissue, blood, and cyst fluid motion, in-vivo,” 2003 IEEE Symposium on Ultrasonics, Vol. 1, pp. 841-846, 2003.
[6]C. H. Choi, W. Chang, S. Y. Lee, “Blind source separation of speech and music signals using harmonic frequency dependent independent vector analysis,” Electronics Letters, Vol. 48, No. 2, pp. 124-125, 2012.
[7]J. Anemüller, T. J. Sejnowski, S. Makeig, “Complex Independent Component Analysis of Frequency-Domain Electroencephalographic Data,” Neural Networks, Vol. 16, No. 9, pp. 1311-1323, 2003.
[8]N. Ono, “Stable and Fast Update Rules for Independent Vector Analysis Based on Auxiliary Function Technique,” Proc. of IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, pp. 189-192, 2011.
[9]W. C. Huang, S. H. Hung, J. F. Chung, M. H. Chang, L. D. Van, C. T. Lin, “FPGA Implementation of 4-Channel ICA for on-line EEG Signal Separation,” Proc. of 2008 Conference on Biomedical Circuits and Systems, pp. 65-68, 2008.
[10]A. J. Bell, T. J. Sejnowski, “An Information-Maximization Approach to Blind Separation and Blind Deconvolution,” Neural Computation, Vol. 7, No. 6, pp.1129-1159, 1995.
[11]A. L. Price, N. J. Patterson, R. M. Plenge, M. E. Weinblatt, N. A. Shadick, D. Reich, ” Principal Components Analysis Corrects for Stratification in Genome-Wide Association Studies,” Nature Genetics, Vol. 38, No. 8, pp. 904, 2006.
[12]王小川(2008)。《語音訊號處理》。臺北：全華圖書。
[13]Jyh-Shing Roger Jang, "Audio Signal Processing and Recognition," available at the links for on-line courses at the author's homepage at http://www.cs.nthu.edu.tw/~jang.
[14]K. Pearson, “On Lines and Planes of Closest Fit to Systems of Points in Space,” Dublin Philosophical Magazine and Journal of Science, Vol. 2, No. 11, pp. 559-572, 1901.
[15]Y. LeCun, Y. Bengio, G. Hinton, “Deep Learning,” Nature, Vol. 521, No. 7553, pp.436, 2015.
[16]J. W. Cooley, J. W. Tukey, “An Algorithm for the Machine Calculation of Complex Fourier Series,” Mathematics of Computation, Vol. 19, No. 90, pp. 297-301, 1965.
[17]E. Keogh, C. A. Ratanamahatana, “Exact Indexing of Dynamic Time Warping,” Knowledge and Information Systems, Vol. 7, No. 3, pp. 358-386, 2005.
[18]Z. Wang, A. C. Bovik, H. R. Sheikh, E. P. Simoncelli, “Image Quality Assessment: from Error Visibility to Structural Similarity,” IEEE Transactions on Image Processing, Vol. 13, No .4, pp. 600-612, 2004.
[19]C. Y. Chang, Y. C. Hsiao, S. T. Chen, “Application of Incremental SVM Learning for Infant Cries Recognition,” Proc. of the 18th International Conference In Network-Based Information Systems, pp. 607-610, 2015.
[20]R. Cohen, Y. Lavner, “Infant Cry Analysis and Detection,” Proc. of 2012 IEEE 27th Convention on Electrical & Electronics Engineers in Israel (IEEEI), pp. 1-5, 2012.
[21]Y. L. Len, “An Investigation of Frequency Domain ICA for Speech Signal Separation,” National Chiao Tung University Master Thesis, 2004.
[22]X. C. Wang, “A Speed-up Joint Approximate Diagonalize Method for Convolutive Mixed Blind Source Separation,” National Tsing Hua University Master Thesis, 2008.