一种基于脉冲耦合神经网络的语音情感识别新方法

针对汉语语音情感识别问题,提出了一种基于脉冲耦合神经网络(PCNN)的识别方法。该方法将语音转化为语谱图后输入到PCNN,得到输出图像的神经元点火序列及其熵序列作为语音情感的特征,利用其特征实现语音情感识别。实验结果表明,该方法可以有效地识别“高兴”与“平常”这两种不同的情感。该方法将PCNN引入到语音情感识别的应用研究中,开拓了语音和图像信号结合处理的新领域,同时对于PCNN的理论研究和实际应用具有重要的现实意义。     

基于Attention-CGRU网络的中文语音情感识别

正确识别语音中包含的情感信息可以大幅提高人机交互的效率.目前,语音情感识别系统主要由语音特征抽取和语音特征分类两步组成.为了提高语音情感识别准确率,选用语谱图而非传统声学特征作为模型输入,采用基于attention机制的CGRU网络提取语谱图中包含的频域信息和时域信息.实验结果表明:在模型中引入注意力机制有利于减少冗余信息的干扰,并且相较于基于LSTM网络的模型,采用GRU网络的模型预测精确度更高,且在训练时收敛更快,与基于LSTM的基线模型相比,基于GRU网络的模型训练时长只有前者的60%.     

基于频谱增强和卷积宽度学习的音乐流派分类

针对频谱图对于音乐特征挖掘较弱、深度学习分类模型复杂且训练时间长的问题,设计了一种基于频谱增强和卷积宽度学习(CNNBLS)的音乐流派分类模型.该模型首先通过SpecAugment中随机屏蔽部分频率信道的方法增强梅尔频谱图,再将切割后的梅尔频谱图作为CNNBLS的输入,同时将指数线性单元函数(ELU)融合进CNNBLS的卷积层,以增强其分类精度.相较于其他机器学习网络框架, CNNBLS能用少量的训练时间获得较高的分类精度.此外, CNNBLS可以对增量数据进行快速学习.实验结果表明:无增量模型CNNBLS在训练400首音乐数据可获得90.06%的分类准确率,增量模型Incremental-CNNBLS在增加400首训练数据后可达91.53%的分类准确率.     

基于NIR技术的PET/PVC废旧塑料分离系统

介绍了一种基于近红外探测技术的PET/PVC废旧塑料分离系统.该系统根据不同塑料对近红外光谱吸收不同的原理,结合窄带滤光片和光电传感器设计了红外检测装置,开发了以STM32F103X为核心,包括信号处理、A/D转换、控制模块、分离模块、CAN总线通信模块在内的自动鉴别分离系统.实验结果表明:该系统测量结果准确,响应速度...     

Deep Scalogram Representations for Acoustic Scene Classification

Spectrogram representations of acoustic scenes have achieved competitive performance for acoustic scene classification. Yet, the spectrogram alone does not take into account a substantial amount of time-frequency information. In this study, we present an approach for exploring the benefits of deep scalogram representations, extracted in segments from an audio stream. The approach presented firstly transforms the segmented acoustic scenes into bump and morse scalograms, as well as spectrograms; secondly, the spectrograms or scalograms are sent into pre-trained convolutional neural networks; thirdly, the features extracted from a subsequent fully connected layer are fed into (bidirectional) gated recurrent neural networks, which are followed by a single highway layer and a softmax layer; finally, predictions from these three systems are fused by a margin sampling value strategy. We then evaluate the proposed approach using the acoustic scene classification data set of 2017 IEEE AASP Challenge on Detection and Classification of Acoustic Scenes and Events (DCASE). On the evaluation set, an accuracy of 64.0% from bidirectional gated recurrent neural networks is obtained when fusing the spectrogram and the bump scalogram, which is an improvement on the 61.0% baseline result provided by the DCASE 2017 organisers. This result shows that extracted bump scalograms are capable of improving the classification accuracy, when fusing with a spectrogram-based system.      

Axlebox accelerations: Their acquisition and time–frequency characterisation for railway track monitoring purposes

Railway track maintenance is becoming a real challenge for Railway Engineers due to the need of meeting increasingly high quality requirements by means of cost-effective procedures. Frequently, this can be only achieved by implementing some technological developments from other fields into the railway sector, such as Digital Signal Processing. Indeed, the present work delves into data acquisition and processing techniques in order to enhance track surveying processes. For this purpose, run tests on the Metropolitan Rail Network of Valencia (Spain) were carried out, and axlebox accelerations were gathered and analysed in different ways. The results determined the optimal sampling and filtering frequencies as well as the location of accelerometers along the train. Furthermore, by means of spectral analysis and time–frequency representations, diverse track defects, track singularities and vibration modes can be clearly identified. It is shown how, with a Hamming time window of 0.5 s and an overlapping of 95%, a wide set of track defects can be detected, without the need of complementary analyses. These values yield the best results as they are a good compromise between time and frequency resolution and allow for appropriate pattern recognition of the corresponding track singularities and resonant frequencies.     

气质联用仪在气体分析中的应用

通过实验论述了气质联用在气体分析中的应用,对硫化氢、氧硫化碳气体的线性、灵敏度和数据重复性进行了考察,探讨了气质联用仪用于气体定量分析的可靠性,并给出了其它气体的色谱分析条件及谱图。     

并条机振动产生的周期性疵点在波谱图上的响应

本文以分析波谱图为主要手段,测试了细纱、并条所纺纱条的均匀度波谱图,用重复试验的方法绘制了波谱图峰值变化曲线。用 L-TEST 测得相应片段间的均匀度曲线图,进而讨论了因并条机振动而造成的周期性疵点在波谱图、均匀度曲线图上所反映的规律和特点,以及获得该图的测试方法。     

黄台电厂8号机组凝结水升压泵振动分析与处理

介绍了对黄台电厂8号机组乙凝结水升压泵振动的分析和处理过程。振动主要由于基础松动引起,对松动的基础重新浇灌了二次灌浆层后,泵组消振达到标准要求。     

国纺Ⅰ型毛纺智能条粗条干均匀度测试仪的使用实践

介绍了国纺Ⅰ型毛纺智能条粗条干均匀度仪的技术特点，以及其在毛纺厂中使用的实际效果。同时将该仪器测试的结果与乌斯特3型条干仪测试的结果进行了比较，验证了仪器的精确性和实用性。该仪器可根据工厂不同机型配备专家分析软件，应用质量和工作效率能符合工厂实际要求，并将测试结果与乌斯特2001年统计值进行了对比分析.