Binaural localization algorithm based on deep learning
-
摘要: 针对多种定位因素存在复杂关联且不易准确提取的问题，提出了以完整双耳声信号作为输入的、基于深度学习的双耳声源定位算法。首先，分别采用深层全连接后向传播神经网络(Deep Back Propagation Neural Network，D-BPNN)和卷积神经网络(Convolutional Neural Network， CNN)实现深度学习框架；然后，分别以水平面 15°、30°和 45°空间角度间隔的双耳声信号进行模型训练；最后，采用前后混乱率、定位准确率与训练时长等指标进行算法有效性分析。模型预测结果表明，CNN模型的前后混乱率远低于 D-BPNN；D-BPNN模型的定位准确率能够达到87%以上，而 CNN模型的定位准确率能够达到 98%左右；在相同实验条件下，CNN模型的训练时长大于 D-BPNN，且随着水平面角度间隔的减小，两者训练时长之间的差异愈发显著。Abstract: Due to existence of complicated relationships between multiple localization cues, which causes them hard to be extracted accurately, a deep learning-based binaural sound source localization algorithm with complete binaural sound signals as input is proposed. Firstly, the deep fully connected back propagation neural network (D-BPNN) and the convolutional neural network (CNN) are used to implement the deep learning framework respectively. And then, binaural sound source signals with uniform azimuthal spacing of 15°, 30° and 45° in horizontal plane are applied to model training respectively. Finally, indicators such as front-back confusion rate, localization accuracy and training duration are used to investigate effectiveness of the models. The model prediction results show that the front-back confusion rate of the CNN model is much lower than that of D-BPNN model. The localization accuracy of the DBPNN model can reach more than 87%, while the localization accuracy of the CNN model is about 98%. Under the same experimental conditions, the training time of CNN model is longer than that of D-BPNN model; Moreover, this difference in training time becomes more and more obviously as the azimuthal spacing in the horizontal plane decreases.
-
-
[1] BLAUERT J. Spatial hearing:the psychophysics of human sound localization[M]. Cambridge:MIT Press, 1997.
[2] MIDDLEBROOKS J C. Sound localization[J]. Handbook of Clinical Neurology, 2015, 129:99-116.
[3] ALTMANN C F, UEDA R, BUCHER B, et al. Trading of dynamic interaural time and level difference cues and its effect on the auditory motion-onset response measured with electroencephalography[J]. NeuroImage, 2017, 159:185-194.
[4] BAUMGARTNER R, MAJDAK P, LABACK B. Modeling sound-source localization in sagittal planes for human listeners[J]. The Journal of the Acoustical Society of America, 2014, 136(2):791-802.
[5] LANGENDIJK E H A, BRONKHORST A W. Contribution of spectral cues to human sound localization[J]. The Journal of the Acoustical Society of America, 2002, 112(4):1583-1596.
[6] BIANCO M J, GERSTOFT P, TRAER J, et al. Machine learning in acoustics:theory and applications[J]. The Journal of the Acoustical Society of America, 2019, 146(5):3590.
[7] GILL D, TROYANSKY L, NELKEN I. Auditory localization using direction-dependent spectral information[J]. Neurocomputing, 2000, 32-33:767-773.
[8] CHUNG W, CARLILE S, LEONG P. A performance adequate computational model for auditory localization[J]. The Journal of the Acoustical Society of America, 2000, 107(1):432-445.
[9] JIN C, SCHENKEL M, CARLILE S. Neural system identification model of human sound localization[J]. The Journal of the Acoustical Society of America, 2000, 108(3 Pt 1):1215-1235.
[10] HINTON G E, OSINDERO S, TEH Y W. A fast learning algorithm for deep belief nets[J]. Neural Computation, 2006, 18(7):1527-1554.
[11] 倪俊帅,赵梅,胡长青.基于深度学习的舰船辐射噪声多特征融合分类[J].声学技术, 2020, 39(3):366-371. NI Junshuai, ZHAO Mei, HU Changqing. Multi-feature fusion classification of ship radiated noise based on deep learning[J]. Technical Acoustics, 2020, 39(3):366-371.
[12] 罗笑雪,柯雨璇,郑成诗,等.联合谱和空间特征的深度学习语音增强研究[J].声学技术, 2019, 38(5):467-468. LUO Xiaoxue, KE Yuxua, ZHENG Chengshi, et al. Deep learning-based speech enhancement using both spectral and spatial features[J]. Technical Acoustics, 2019, 38(5):467-468.
[13] 丁建策,厉剑,彭任华,等.室内两步法监督式学习双耳声源距离估计[J].声学学报, 2019, 44(4):405-416. DING Jiance, LI Jian, PENG Renhua, et al. Two-stage supervised binaural distance estimation in room environments[J]. Acta Acustica, 2019, 44(4):405-416.
[14] DING J C, KE Y X, CHENG L J, et al. Joint estimation of binaural distance and azimuth by exploiting deep neural networks[J]. The Journal of the Acoustical Society of America, 2020, 147(4):2625.
[15] GAVIRIA J F, ESCALANTE-PEREZ A, CASTIBLANCO J C, et al. Deep learning-based portable device for audio distress signal recognition in urban areas[J]. Applied Sciences, 2020, 10(21):7448.
[16] GAMPER H, TASHEV I J. Blind reverberation time estimation using a convolutional neural network[C]//2018 16th International Workshop on Acoustic Signal Enhancement (IWAENC). Tokyo, Japan. IEEE, 2018:136-140.
[17] CHEN R, SCHMIDT H. Model-based convolutional neural network approach to underwater source-range estimation[J]. The Journal of the Acoustical Society of America, 2021, 149(1):405-420.
[18] NEILSEN T B, ESCOBAR-AMADO C D, ACREE M C, et al. Learning location and seabed type from a moving mid-frequency source[J]. The Journal of the Acoustical Society of America, 2021, 149(1):692.
[19] GARDNER W G, MARTIN K D. HRTF measurements of a KEMAR[J]. The Journal of the Acoustical Society of America, 1995, 97(6):3907-3908.
[20] 谢菠荪.头相关传输函数与虚拟听觉[M].北京:国防工业出版社, 2008. -
期刊类型引用(2)
1. 梅鹏程，杨吉斌，张强，黄翔. 一种基于三维卷积的声学事件联合估计方法. 计算机科学. 2023(03): 191-198 . 
百度学术
2. 卢炽华，薛齐凡，刘志恩，朱亚伟，彭文杰，李放. 基于SincNet增强的时延估计声源定位算法研究. 武汉理工大学学报. 2023(10): 127-134 . 百度学术
其他类型引用(4)
计量
- 文章访问数: 848
- HTML全文浏览量: 0
- PDF下载量: 1205
- 被引次数: 6
下载:
