Deep Learning-based Environmental Sound Classification Using Feature Fusion and Data Enhancement

Environmental sound classification(ESC)involves the process of distinguishing an audio stream associated with numerous environmental sounds.Some common aspects such as the framework difference,overlapping of different sound events,and the presence of various sound sources during recording make the ESC task much more complicated and complex.This research is to propose a deep learning model to improve the recognition rate of environmental sounds and reduce the model training time under limited computation resources.In this research,the performance of transformer and convolutional neural networks(CNN)are investigated.Seven audio features,chromagram,Mel-spectrogram,tonnetz,Mel-Frequency Cepstral Coefficients(MFCCs),delta MFCCs,delta-delta MFCCs and spectral contrast,are extracted fromtheUrbanSound8K,ESC-50,and ESC-10,databases.Moreover,this research also employed three data enhancement methods,namely,white noise,pitch tuning,and time stretch to reduce the risk of overfitting issue due to the limited audio clips.The evaluation of various experiments demonstrates that the best performance was achieved by the proposed transformer model using seven audio features on enhanced database.For UrbanSound8K,ESC-50,and ESC-10,the highest attained accuracies are 0.98,0.94,and 0.97 respectively.The experimental results reveal that the proposed technique can achieve the best performance for ESC problems.

Comparative Analysis of Different Sampling Rates on Environmental Sound Classification Using the Urbansound8k Dataset

Environmental sound classification (ESC) has gained increasing attention in recent years. This study focuses on the evaluation of the popular public dataset Urbansound8k (Us8k) at different sampling rates using hand crafted features. The Us8k dataset contains environment sounds recorded at various sampling rates, and previous ESC works have uniformly resampled the dataset. Some previous work converted this data to different sampling rates for various reasons. Some of them chose to convert the rest of the dataset to 44,100, as the majority of the Us8k files were already at that sampling rate. On the other hand, some researchers down sampled the dataset to 8000, as it reduced computational complexity, while others resampled it to 16,000, aiming to achieve a balance between higher classification accuracy and lower computational complexity. In this research, we assessed the performance of ESC tasks using sampling rates of 8000 Hz, 16,000 Hz, and 44,100 Hz by extracting the hand crafted features Mel frequency cepstral coefficient (MFCC), gamma tone cepstral coefficients (GTCC), and Mel Spectrogram (MelSpec). The results indicated that there was no significant difference in the classification accuracy among the three tested sampling rates.

Evaluation of Individual and Environmental Sound Pressure Level and Drawing Noise-Isosonic Maps Using Surfer V.14 and Noise at Work V.5.0

Noise pollution is one of the common physical harmful factors in many work environments.The current study aimed to assess personal and environmental sound pressure level and project the sound-Isosonic map in one of the Razavi Khorasan Paste manufacture using Surfer V.14 and Noise at work V.5.0.This cross-sectional,descrip-tive study is analytical that was conducted in 2018 in the Paste factory that contains Canister,production and Brewing unit.Following ISO 9612:2009,Casella Cel-320 was used to measure personal sound pressure level,while CEL-450 sound level meter(manufactured by Casella-Cel,the UK)was employed to assess environmental sound pressure level.Statistical analyzes was done using SPSS V.18 and Linear Regression test.The sound-isosonic maps were projected using Surfer V.14 and Noise at work V.5.0.The results of assessing personal sound pressure level indicated that the highest received dose(172.21%)and personal equivalent sound level(87.36 dBA)were recorded for workers in the Canister unit.According to results of measuring of the environmental sound pressure level,out of 16 measurement stations in this unit,overall 87.5%were regarded as danger and caution areas.The lowest and highest sound pressure levels in this units were 61 dBA and 92 dBA that belong to Brewing and Canister units respectively.Results indicate Over 75%of the Canister and production units had a sound pressure level greater than 85 dBA and these two units were regarded as the most dangerous area in terms of noise pollution.It is there-fore necessary to implement noise control measures,apply hearing protection program and auditory tests among workers in these units.

Study of Sound Environment Influenced by the Crowd in Waiting Areas in General Hospitals

In this study,the crowd has been investigated and analyzed in waiting areas in large general hospitals in China in order to find the rules the variations of sound environment with the change of crowd. The field investigation,questionnaire,field-testing and computer simulation have been adopted. The results show that: the social /demographic characteristics of staff and patients are not significantly related to the satisfaction evaluation of sound environment; there is a significant correlation between the population density and LAeq of the background noise in waiting areas; when population density is 0,the LAeq of background noise is not 0 in waiting areas; the loudspeaker should be set in the waiting areas. Loudspeaker arrangements should be integrated into the ceiling lamp or construct facilities along the depth direction of the layout,and the two adjacent speakers recommended distance should be controlled at about 4 m. If the population density is controlled in the reasonable range,and sound absorption,noise reduction processing and electronic queuing system are adopted,sound environment of waiting areas will be built with noise interference relatively small in different population densities.

Effect of ocean environmental factors on sound absorption by boric acid relaxation in sea water

By using the expressions for the maximum absorption per wavelength (αλ),and the relaxation frequency fr of the boric acid relaxation derived previously by the author and employing the related oceanographic literatures, the effects of pressure, temperature, pH and salinity on (αλ)r and ∫r of the boric acid relaxation in sea water have been estimated. Results show that ( αλ), not only increases with pH but also increases approximately linearly with pressure and temperature, and is nearly proportional to the 1. 35 power of salinity. However, pressure, pH and salinity have negligible effect on ∫r; therefore, ∫r, can be approximately expressed as a function of temperature only. Comparisons of the predicted with the measured ( αλ)r and ∫r in different ocean areas are given.

Sound-Environment Monitoring Method Based on Computational Auditory Scene Analysis

Monitoring techniques are a key technology for examining the conditions in various scenarios, e.g., structural conditions, weather conditions, and disasters. In order to understand such scenarios, the appropriate extraction of their features from observation data is important. This paper proposes a monitoring method that allows sound environments to be expressed as a sound pattern. To this end, the concept of synesthesia is exploited. That is, the keys, tones, and pitches of the monitored sound are expressed using the three elements of color, that is, the hue, saturation, and brightness, respectively. In this paper, it is assumed that the hue, saturation, and brightness can be detected from the chromagram, sonogram, and sound spectrogram, respectively, based on a previous synesthesia experiment. Then, the sound pattern can be drawn using color, yielding a “painted sound map.” The usefulness of the proposed monitoring technique is verified using environmental sound data observed at a galleria.

Environmental Sound Recognition Using Double-Level Energy Detection

The performance of classic Mel-frequency cepstral coefficients (MFCC) is unsatisfactory in noisy environment with different sound sources from nature. In this paper, a classification approach of the ecological environmental sounds using the double-level energy detection (DED) was presented. The DED was used to detect the existence of the sound signals under noise conditions. In addition, MFCC features from the frames which were detected the presence of the sound signals by DED were extracted. Experimental results show that the proposed technology has better noise immunity than classic MFCC, and also outperforms time-domain energy detection (TED) and frequency-domain energy detection (FED) respectively.

Create an International Environment for the Peaceful and Steady Development to Boost the Sound Development of Human Rights on the International Stage——An Address at the Opening Ceremony of the 7th Beijing Forum on Human Rights, Spet. 17, 2014

Since its initiation in 2008, the Beijing Forum on Human Rights has been bent on promoting our mutual understanding and expanding our common grounds. It has carried out very fruitful exchanges and researches around many important questions concerning human rights development, and hence attained great achievements. It has played avery constructive role in promoting exchanges about and cooperation in human rights between China and other countries, becoming an important and influential platform for international dialogs over human rights.

结合MGCC特征与多尺度通道注意力的环境声深度学习分类方法

环境声分类技术在家居安全监测、人机语声交互等领域具有关键作用。然而,声源的多样性与混合性给环境声分类方法设计带来了重大挑战。为提高分类准确率与节约计算资源,该文提出一种基于多尺度通道注意力机制的深度学习分类模型。所提模型由特征提取模块、多尺度卷积模块、高效通道注意力模块、输出层四部分组成。首先,通过引入加权型梅尔Gammatone频率倒谱系数(MGCC)挖掘环境声频谱幅值与相位结构信息;其次,融合多尺度卷积核与高效通道注意力机制优选出声频关键局部细节和通道特征;最后,在全连接层采用softmax函数映射特征并输出环境声类型的概率值。所提模型在6种环境声的iFLYTEK、10种环境声的Urbansound8k数据集上开展测试验证,分别取得了94%、76.52%、79.24%(iFLYTEK+Urbansound8k)的分类准确率。消融实验结果进一步表明:引入的多尺度卷积模块、通道注意力机制模块对分类准确率的提升贡献率分别接近于3.77%和1.89%。实验还详细对比了7种现有的深度学习分类方法,所提算法在分类准确率上排名第二;另外,在同级别算法中如ResNet18、GoogLeNet,所提算法在模型参数量和计算复杂度方面上实现了进一步的约减。

一种面向微控制器上环境声音分类的DNN压缩方法

环境声音分类(Environmental Sound Classification,ESC)是非语音音频分类任务最重要的课题之一。近年来,深度神经网络(Deep Neural Network,DNN)方法在ESC方面取得了许多进展。然而,DNN是计算和存储密集型的,无法直接部署到基于微控制器(Microcontroller Unit,MCU)的物联网设备上。针对这一问题,本文提出一种用于资源高度受限设备的DNN压缩方法。由于DNN模型参数规模较大无法直接部署,因此提出使用剪枝方法进行大幅压缩,并针对该操作带来的精度损失问题,设计一种基于模型中间层特征信息的知识蒸馏方法。基于STM32F746ZG设备在公开的数据集(UrbanSound8K、ESC-50)上进行测试,实验结果表明,本文方法能够获得高达97%的压缩率,同时保持良好的推理精度和速度。

融合Swin Transformer和CNN的环境声音分类模型

环境声音分类已经成为计算机听觉领域的一项重要任务,可以作为计算机视觉的补充,帮助设备更好地理解环境和用户需求,具有广泛的应用前景,将对人类生活产生积极影响。近年来,环境声音分类领域采用了具有自注意力机制的Transformer模型,然而现有模型需要较大的内存,同时依赖于预训练的视觉模型,无法较好提取音频特征。为了解决这些问题并提高环境声音分类准确度,提出了一种新的具有双分支结构的Swin Conformer环境声音分类模型。通过融合卷积神经网络和具有窗口自注意力机制的Swin Transformer模型,以交互方式融合双分支特征并引入令牌语义模块。结果表明:Swin Conformer模型在ESC-50和UrbanSound8K公共数据集上分别通过验证实现了98.1%和96.8%的分类准确度。与现有模型相比,具有更高的分类准确度,证明了该模型在环境声音分类任务中的可行性和优越性。

基于并联型神经网络的环境声音分类

针对传统单输入模型在环境声音分类中准确率不高的问题,提出一种基于时域特征和频域特征并联型特征融合神经网络。在该网络中,首先通过数据增强的方法来处理原始音频;其次处理后的原始音频数据和梅尔(Mel)频谱特征数据分别送入原始波形网络和Mel频谱网络,得到其时域和频谱特征后,进行特征融合;最后,将特征融合后的结果送入SoftMax分类器进行分类。本文在UrbanSound8K数据集上进行了实验验证,最终分类准确率高达96.03%,优于其他模型。

徐州市近十年区域声环境质量变化趋势分析及防治策略

为综合评价徐州市区域声环境质量状况,进一步探讨徐州市区域声环境存在的问题,基于徐州市2013年~2022年区域声环境监测数据,综合分析了徐州市区域声环境质量现状及变化趋势。结果表明,近十年徐州市区域声环境质量处于53.5~56.7 dB(A)之间,总体保持稳定且有向好趋势;从声源构成来看,社会生活噪声在历年城市区域环境噪声的构成占比巨大并且逐年上升,社会生活噪声来源广泛不易控制;交通噪声在历年非安静区所覆盖面积呈现逐步上升的趋势,要使区域声环境质量得到有效的改善,交通噪声源的控制显得尤为重要。

硅溶胶改性的绿色环保室内吸声材料制备与性能研究

为同时达到缓解噪音污染和环境污染的目的,利用针叶木纤维、硅溶胶、苯丙乳液、十二烷基苯磺酸钠、发泡浆料等原材料制备了一种室内绿色环保可降解吸声材料,并对材料的吸声性能进行了研究。结果表明,利用硅溶胶对吸声材料进行改性,可显著提升材料的力学性能和吸声性能,当硅溶胶掺量为8%时,可达到材料孔隙结构和吸声效果的最优平衡;增大植物纤维密度和厚度,可以提升材料在中低频率下的吸声系数,但对于中高频率的吸声性能提升效果并不显著;Delany-Bazley模型能够较为准确地预测不同流阻率植物纤维基吸声材料的吸声系数,预测值与试验数据平均误差仅为1.8%。

基于黄金尾矿处理与综合利用研究的文献计量分析

黄金尾矿中含有较多的重金属元素和有价元素,若处置不当不仅浪费资源,而且污染环境、危害人类健康,存在巨大的安全隐患。利用文献计量学和VOS viewer软件,针对国内外黄金尾矿的处理和综合利用研究态势开展研究,并对未来研究热点进行预测。结果显示:黄金尾矿处理及综合利用研究呈现增长趋势,各国的研究内容密切相关;研究内容主要包括黄金尾矿浮选、黄金尾矿综合利用以及黄金尾矿库的生态修复。展望了黄金尾矿未来无害化处理及高值化利用的发展方向,为黄金尾矿的综合利用提供参考依据。

福州市五城区声环境质量状况及污染防治对策研究

以“十三五”期间福建省福州市五城区声环境质量监测数据及污染投诉情况为基础,从功能区声环境质量、区域声环境质量、道路交通声环境质量、噪声投诉情况等方面对现状进行分析,基于分析结果剖析了福州市在快速发展过程中存在的社会生活噪声污染日益突出、施工噪声持续扰民、交通噪声夜间污染严重、声环境监测网络亟须完善、噪声权责监管体系不清等问题,并提出应加强噪声源头管控、合理开展规划布局,建立噪声区划动态调整机制、健全声环境质量监测网络,坚持分类管控、实施精准科学依法治污,健全污染防治制度体系、厘清部门监管职责,完善联合执法协调联动机制、部门齐抓共管等污染防治对策。

河源市前汛期暖区暴雨环境参数的统计分析

采用2007—2022年4—6月河源市探空资料和地面常规观测资料,统计分析了河源市前汛期暖区暴雨时空分布特征和环境参数,并基于百分位法提取了不同等级暖区暴雨的预报关键阈值。结果表明:4—6月暖区暴雨事件逐月明显增加,且72.5%出现在南海夏季风爆发以后;发生频次的空间分布自西向东逐渐减少,高频次站点均处于东北-西南向的喇叭口地形附近;925与850 hPa温度露点差的平均值、500 hPa温度露点差、整层比湿积分、抬升指数、K指数和0~3 km垂直风切变对典型河源暖区暴雨事件的发生具有较清晰指示意义。

基于嵌入式机器学习的智能家居安全预警系统设计

为满足与日俱增的智能家居安全管理需求,文章设计并实现了一套智能家居安全预警系统,包括监测终端节点和Web客户端软件。监测终端节点在开源硬件Arduino Nano 33 BLE上部署改进的卷积神经网络模型,通过采集家居环境声音判断是否有异常事件发生。Web客户端软件实时显示异常事件,及时向用户发送预警邮件。实验结果表明,该家居安全预警系统能够有效监测家居环境安全,满足用户对家居安全保障的现实需求。

公路项目环境影响评价中声屏障尺寸研究

通过对线源设置声屏障后插入损失的计算方法的研究,明确了公路项目设计降噪目标值的确定方法,推导出了声屏障尺寸计算公式,并通过实例进行了验算,计算结果与现行规范具有较好的适用性,可以作为环境影响评价过程中确定声屏障尺寸的依据。