An Improved Voice Activation Detection Method Based on Energy Acceleration Parameters and Support Vector Machine

Abstract:

Voice activation detection is a very important part in speech related domain. The classic voice activation detection normally depends on feature parameters in time or frequency domain, or parameters from statistical model. An improved voice activation detection method based on energy acceleration parameters and support vector machine is proposed in this paper. The energy acceleration parameters are the voice activation detection parameters in ETSI Advanced front-end feature extraction algorithm. The training period of support vector machine is based on energy acceleration parameters and manually appended class labels of each frame. In the detection period, the detection result is derived from energy acceleration parameters and Lagrange parameters calculated from training period. The experimental result shows that the false alarm rate of proposed method is greatly decreased. It has been observed that the voice activation detection proposed is better than the voice activation detection in ETSI Advanced front-end feature extraction algorithm.

• Periodical: Advanced Materials Research (Volume 981)
• Pages: 287-291
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.981.287
• Online since: July 2014
• Authors: Qian Liu*, Jin Xiang Wang, Ming Jiang Wang, Pan Pan Jiang
• Keywords: Energy Acceleration Parameter, Support Vector Machine (SVM), Voice Activation Detection

References

[1] Jamal Saeedi · Seyed Mohammad Ahadi · Karim Faez Robust voice activity detection directed by noise classification SIViP, April, (2013).

DOI: 10.1007/s11760-013-0479-5

[2] Beritelli, F., Casale, S., Ruggeri, G.: Performance evaluation and comparison of ITU-T/ETSI voice activity detectors. In: Proceedings ICASSP, (2001) , p.1425–1428.

DOI: 10.1109/icassp.2001.941197

[3] Srinivasant, K., Gersho, A.: Voice activity detection for cellular networks. In: Proceedings IEEE Speech Coding, Workshop, ( 1993), p.85–86.

[4] SOHN J., KIM N.S., SUNG W.: A statistical model-based voice activity detection, IEEE Signal Process. Lett., 6, (1), (1999), p.1–3.

DOI: 10.1109/97.736233

[5] ITU-T Recommendation G729-Annex B, November, (1996).

[6] Speech processing, transmission and quality aspects(STQ); Distributed speech recognition; Advanced front-end feature extraction algorithm; compression algorithms, ETSI ES 202, (2007).

[7] Q. -H. Jo, J. -H. Chang J.W. Shin N.S. Kim. Statistical model-based voice activity detection using support vector machine, IET Signal Processing, Vol. 3, Iss. 3, (2009), p.205–210.

DOI: 10.1049/iet-spr.2008.0128

[8] Tomi Kinnunen, Evgenia Chernenko, Marko Tuononen, Pasi Fränti, Haizhou Li, Voice Activity Detection Using MFCC Features and Support Vector Machine, International Conference on SPECOM, Vol. 2, (2007), pp.556-561.

[9] ENQING D., GUIZHONG L., YATONG Z., XIAODI Z.: Applying support vector machines to voice activity detection. Proc. Int. Conf. Signal Process., vol. 2, (2002), p.1124–1127.

DOI: 10.1109/icosp.2002.1179987

[10] Tom Fawcett. ROC Graphs: notes and practical considerations for researchers. HP laboratories, (2004).