화자 식별에서의 배경화자데이터를 이용한 히스토그램 등화 기법
Histogram Equalization Using Background Speakers’ Utterances for Speaker Identification
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
In this paper, we propose a novel approach to improve histogram equalization for speaker identification. Our method collects all speech features of UBM training data to make a reference distribution. The ranks of the feature vectors are calculated in the sorted list of the collection of the UBM training data and the test data. We use the ranks to perform order-based histogram equalization. The proposed method improves the accuracy of the speaker recognition system with short utterances. We use four kinds of speech databases to evaluate the proposed speaker recognition system and compare the system with cepstral mean normalization (CMN), mean and variance normalization (MVN), and histogram equalization (HEQ). Our system reduced the relative error rate by 33.3% from the baseline system.
Abstract
In this paper, we propose a novel approach to improve histogram equalization for speaker identification. Our method collects all speech features of UBM training data to make a reference distribution. The ranks of the feature vectors are calculated in the sorted list of the collection of the UBM training data and the test data. We use the ranks to perform order-based histogram equalization. The proposed method improves the accuracy of the speaker recognition system with short utterances. We use four kinds of speech databases to evaluate the proposed speaker recognition system and compare the system with cepstral mean normalization (CMN), mean and variance normalization (MVN), and histogram equalization (HEQ). Our system reduced the relative error rate by 33.3% from the baseline system.
참고문헌 (12)
- Atal. B. S, "Effectiveness of linear predictioncharacteristics of the speech wave for automatic speakeridentification and verification", Journal of the Acoustical Societyof America, 55(6), pp.1304~1312 (1974)
- Viikki. O., "Cepstral domain segmentalfeature vector normalization for noise robust speechrecognition", Speech Communication, 25, pp.133~147 (1998)
- Huang. X., "Spoken LanguageLanguage Processing: A Guide to Theory, Algorithm, andSystem Development" (2001), Prentice-Hall
- Moreno. P. J., "A vectorTaylor series approach for environment independent speechrecognition", Proc. ICASSP, pp.733~736 (1996)
- Kim. N. S, "Statistical linear approximation forenvironment compensation", IEEE Signal Processing Letters, 5(1), pp.8~10 (2008)
- Gonzalez. R. C., "Digital ImageProcessing" (1987), Addision-Wesley
- Pelecanos. J., "Feature warping forrobust speaker verification", A Speaker Odyssey - The speakerrecognition workshop, pp.213~218 (2001)
- Skosan. M., "Matching featuredistributions for robust speaker verification", Proc. PRASA, pp.42~47 (2004)
- Skosan. M., "Modified segmentalhistogram equalization for robust speaker verification", Pattern Recognition Letters, 27(5), pp.479~486 (2006)
- Segura. J. C., "Cpestral domain segmental nonlinearfeature transformations for robust speech recognition", IEEESignal Processing Letters, 11, pp.517~520 (2006)
- Torre. A. de la, "Histogramequalization of speech representation for robust speechrecognition", IEEE Trans. Speech and Audio Processi, 13, pp.355~366 (2005)
- Reynolds. D. A., "Speaker verification using adapted gaussian mixture models", Digital Signal Processing, 10, pp.19~41 (2000)