Speech Endpoint Detection Based on Fractal Dimension with Adaptive Threshold

doi:10.12068/j.issn.1005-3026.2020.01.002

Abstract

Abstract: Considering the limitation of fixed threshold method in speech endpoint detection， in order to improve the robustness and accuracy of speech endpoint detection under low SNR(signal noise ratio)， a novel speech detection algorithm was proposed based on adaptive threshold in fractal dimension. By analyzing the mechanism of speech signal generation， the fractal dimension was applied to the detection of speech starting and ending points， and an adaptive threshold was designed to avoid noise interference and to achieve real-time detection. The simulation results show that， compared with the traditional short-term energy detection algorithm， the proposed algorithm can effectively realize the endpoint detection of noisy speech under the low SNR， and has better robustness to noise interference.

Key words: speech endpoint detection, fractal dimension, adaptive threshold, low SNR(signal noise ratio), robustness

CLC Number:

TN912.3

ZHENG Yan， GAO Shuang. Speech Endpoint Detection Based on Fractal Dimension with Adaptive Threshold[J]. Journal of Northeastern University Natural Science, 2020, 41(1): 7-11.

References

[1]Seman N，Bakar Z A，Bakar N A.An evaluation of endpoint detection measures for Malay speech recognition of an isolated words［J］.Information Technology，2010(10):1628-1635.
[2]Morita S，Unoki M，Lu X，et al.Robust voice activity detection based on concept of modulation transfer function in noisy reverberant environments［J］.Journal of Signal Processing Systems，2016，82(2):163-173.
[3]Di W U，Zhao H，Huang C，et al.Speech endpoint detection in low-SNRs environment based on perception spectrogram structure boundary parameter［J］.Chinese Journal of Acoustics，2014，39(4):392-399.
[4]李晶皎，安冬，王骄.基于EEMD和ICA的语音去噪方法［J］.东北大学学报(自然科学版)，2011，32(11):1554-1557.(Li Jing-jiao，An Dong，Wang Jiao.Speech denoising method based on the EEMD and ICA approaches［J］.Journal of Northeastern University(Natural Science)，2011，32(11):1554-1557.)
[5]Eshaghi M，Mollaei M R K.Voice activity detection based on using wavelet packet［J］.Digital Signal Processing，2010，20(4):1102-1115.
[6]Lu Y Y，Zhou N，Xiao K，et al.Improved speech endpoint detection algorithm in strong noise environment［J］.Journal of Computer Applications，2014，34(5):1386-1390.
[7]Liang Y S，Su W Y.Fractal dimensions of fractional integral of continuous functions［J］.Acta Mathematica Sinica:English Series，2016，32(12):1494-1508.
[8]刘悦，王晓婷.短时频域分形端点检测算法［J］.微电子与计算机，2015，32(9):82-84.(Liu Yue，Wang Xiao-ting.A speech endpoint detection algorithm based on fractal in short-term frequency domain［J］.Microelectronics and Computer，2015，32(9):82-84.)
[9]Ali Z，Elamvazuthi I，Alsulaiman M，et al.Detection of voice pathology using fractal dimension in a multiresolution analysis of normal and disordered speech signals［J］.Journal of Medical Systems，2016，40(1):20-21.
[10]申希兵，韦容，杨毅.基于频域能量分布的分形维数提取型研究［J］.控制工程，2016，23(6):834-838.(Shen Xi-bing，Wei Rong，Yang Yi.Study on the extraction of fractal dimension based on frequency domain energy distribution［J］.Control Engineering of China，2016，23(6):834-838.)
[11]Jia L，Yin Y，Yang H C.Endpoint detection of noisy speech based on fractal dimension［J］.Journal of Shenyang Aerospace University，2017，34(5):63-67.
[12]张志敏，郭英，王博.一种基于倒谱特征的语音端点检测改进算法［J］.电声技术，2016(4):40-43.(Zhang Zhi-min，Guo Ying，Wang Bo.An improved voice activity detection method based on cepstral features［J］.Voice Technology，2016(4):40-43.)
[13]Sun L，Su M，Yang Z.An adaptive speech endpoint detection method in low SNR environments［J］.International Journal of Speech Technology，2017，20(5):1-8.
[14]Ezeiza A，de Ipia K L，Hernndez C，et al.Enhancing the feature extraction process for automatic speech recognition with fractal dimensions［J］.Cognitive Computation，2013，5(4):545-550.
[15]关守平，房少纯.一种新型的区间-粒子群优化算法［J］.东北大学学报(自然科学版)，2012，33(10):1381-1384.(Guan Shou-ping，Fang Shao-chun.A new particle swarm optimization algorithm［J］.Journal of Northeastern University(Natural Science)，2012，33(10):1381-1384.)