用多频带能量分布检测低信噪比声音事件  

Detection of Sound Event under Low SNR Using Multi-band Power Distribution

作　　者：李应[1] 吴灵菲 LI Ying;WU Lingfei(College of Mathematics and Computer Science,Fuzhou University,Fuzhou 350116,China;Fujian Province Key Laboratory of Information Security of Network Systems,Fuzhou University,Fuzhou 350116,China)

机构地区：[1]福州大学数学与计算机科学学院,福州350116 [2]网络系统信息安全福建省高校重点实验室,福州350116

出　　处：《电子与信息学报》2018年第12期2905-2912,共8页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(61075022);福建省自然科学基金(2018J01793)~~

摘　　要：该文针对低信噪比噪声环境下的声音事件检测问题,提出基于多频带能量分布图离散余弦变换的声音事件检测的方法。首先,将声音数据转化为gammatone频谱,并计算其多频带能量分布;接着,对多频带能量分布图进行8×8分块与离散余弦变换;然后,对8×8的离散余弦变换系数进行Zigzag扫描,抽取离散余弦变换系数的主要系数作为声音事件的特征;最后,利用随机森林分类器对特征建模与检测。实验结果表明,在低信噪比及各种噪声环境下,该文提出的方法具有良好的检测效果。As to the problem of sound event detection in low Signal-Noise-Ratio （SNR） noise environments, a method is proposed based on discrete cosine transform coefficients extracted from multi-band power distribution image. First, by using gammatone spectrogram analysis, sound signal is transformed into multi- band power distribution image. Next, 8×8 size blocking and discrete cosine transform are applied to analyze the multi-band power distribution image. Based on the main Zigzag coefficients which are scanned fiom the discrete cosine transform coefficients, features of sound event are constructed. Finally, features are modeled and detected through random forests classifier. The results show that the proposed method achieves a better detection performance in low SNR comparing to other methods.

关 键 词：声音事件检测 多频带能量分布 随机森林 离散余弦变换 

分 类 号：TP391.42[自动化与计算机技术—计算机应用技术]