Cases are presented to reveal how modern computerised infra-red thermal imaging has the potential to assist in early breast cancer detection. The history of thermography and some recent controversies surrounding mammo...Cases are presented to reveal how modern computerised infra-red thermal imaging has the potential to assist in early breast cancer detection. The history of thermography and some recent controversies surrounding mammography are discussed. Examples of thermal imaging combined with naturopathic interventions are described. Since 2002, more than 8000 women in New Zealand have chosen to include thermal imaging as a part of their breast health management. Breast thermal imaging combined with relevant health advice, resulted in a perceived worthwhile benefit to patients in managing overall health.展开更多
A new method to detect multiple outliers in multivariate data is proposed. It is a combination of minimum subsets, resampling and self-organizing map (SOM) algorithm introduced by Kohonen,which provides a robust way w...A new method to detect multiple outliers in multivariate data is proposed. It is a combination of minimum subsets, resampling and self-organizing map (SOM) algorithm introduced by Kohonen,which provides a robust way with neural network. In this method, the number and organization of the neurons are selected by the characteristics of the spectra, e.g., the spectra data are often changed linearly with the concentration of the components and are often measured repeatedly, etc. So the spatial distribution of the neurons can be arranged by this characteristic. With this method, all the outliers in the spectra can be detected, which cannot be solved by the traditional method, and the speed of computation is higher than that of the traditional neural network method. The results of the simulation and the experiment show that this method is simple, effective, intuitionistic and all the outliers in the spectra can be detected in a short time. It is useful when associated with the regression model in the near infra-red research.展开更多
Alzheimers disease(AD)is a chronic neurodegenerative disease.The symptoms include memoryand spatial learning dificulties,language disorders,and loss of motivation,which get worse overtime,eventually ending in death.No...Alzheimers disease(AD)is a chronic neurodegenerative disease.The symptoms include memoryand spatial learning dificulties,language disorders,and loss of motivation,which get worse overtime,eventually ending in death.No ffective treatments are available for AD,currently.Currenttreatments only attenuate symptoms temporarily and are associated with severe side ffects.Nearinfra-red(NIR)light has been studied for a long time.We investigated the effect of NIR on ADusing a transgenic mouse model,which was obtained by co-injecting two vectors carrying ADmutations in amyloid precursor protein(APP)and presenilin-i(PSEN1)into C57BL/6J mice.The irradiation equipment consisted of an accommodating box and an LED array.The wave-length of NIR light emitted from LED was between 1040 nm and 1090 nm.The power densitydelivered at the level of the mice was approximately 15 mW/cm^(2),Firstly,we treated the micewith NIR for 40 days,Then,the irradiation was suspended for 28 days.Finally,another 15 daystreatment was brought to mice.We conducted Morris water maze and immunofluorescenceanalysis to evaluate the effects of treatment.Immunofuorescence analysis was based on mea-suring the quantity of plaques in mouse brain slices,Our results show that NIR light improvesmemory and spatial learning ability and reduces plaques moderately.NIR light represents apotential treatment for AD.展开更多
In-process damage to a cutting tool degrades the surfacenish of the job shaped by machining and causes a signicantnancial loss.This stimulates the need for Tool Condition Monitoring(TCM)t...In-process damage to a cutting tool degrades the surfacenish of the job shaped by machining and causes a signicantnancial loss.This stimulates the need for Tool Condition Monitoring(TCM)to assist detection of failure before it extends to the worse phase.Machine Learning(ML)based TCM has been extensively explored in the last decade.However,most of the research is now directed toward Deep Learning(DL).The“Deep”formulation,hierarchical compositionality,distributed representation and end-to-end learning of Neural Nets need to be explored to create a generalized TCM framework to perform eciently in a high-noise environment of cross-domain machining.With this motivation,the design of dierent CNN(Convolutional Neural Network)architectures such as AlexNet,ResNet-50,LeNet-5,and VGG-16 is presented in this paper.Real-time spindle vibrations corresponding to healthy and various faulty congurations of milling cutter were acquired.This data was transformed into the time-frequency domain and further processed by proposed architectures in graphical form,i.e.,spectrogram.The model is trained,tested,and validated considering dierent datasets and showcased promising results.展开更多
Introduction: Infra-red (IR) thermometry is a safe and valid method to determine internal and surface temperature in human subjects. Under conditions of brain damage (head injury or stroke) knowledge of changes in the...Introduction: Infra-red (IR) thermometry is a safe and valid method to determine internal and surface temperature in human subjects. Under conditions of brain damage (head injury or stroke) knowledge of changes in the temperature of intracranial tissue is justified because of the vulnerability of neurons to accelerated damage at temperatures at the upper end of the febrile range. Aim: To determine the temperature at the inner canthus (IC) of the eye as a potential surrogate for brain temperature. Methods: Invasive monitoring of deep brain structures, lateral ventricle and deep white matter. IR temperature readings obtained at right and left IC. Results: ?Strong correlations were evident between R and L IC and brain. Close, as well as poor, agreement between?? sites was shown in some patients and at some times. For right hemispheric lesions four had a better correlation between TbrV and TRIC when compared to TLIC.? When the correlation between TbrV and TLIC was better compared to TbrV and TRIC, four had a predominant right hemispheric lesion. Conclusions: Improved techniques for IR thermal imaging accuracy at the bedside has the potential to improve temperature measurement agreement. The predominant lesion side may have a bearing on maximum ipsilateral IC temperature Further studies are ongoing in this pilot study population.展开更多
噪声环境下语音检测准确率偏低是短波通话面临的公开挑战。当前已有方法应用有限,其根源在于难以可靠地在噪音环境下提取准确且高效的语音特征。针对上述问题,提出了一个面向短波通信的低秩方向梯度直方图(Low-rank Histogram of Orient...噪声环境下语音检测准确率偏低是短波通话面临的公开挑战。当前已有方法应用有限,其根源在于难以可靠地在噪音环境下提取准确且高效的语音特征。针对上述问题,提出了一个面向短波通信的低秩方向梯度直方图(Low-rank Histogram of Oriented Gradient,LHOG)话音检测方法。首先,对目标音频源数据进行预处理,实现噪声环境下语音信息的可视化表征;然后,在HOG特征提取器中嵌入低秩化结构,缓解特征中的冗余信息,并降低噪声干扰,从而获得准确且高效的特征;最后,通过常用的SVM分类模型便可在噪声环境中准确快速地区分话音和噪声。测试结果表明,该方法的准确率达到了95.12%,误报率仅为0.96%,漏报率为13.14%。与现有主流方法的对比实验证明,该方法话音检测准确率高,资源占用少,能够有效提高短波通信侦控效率。展开更多
本文提出一种利用双解码卷积循环网络(Dual-decoder Convolutional Recurrent Network,DCRN)代替FxLMS(Filtered-x Least Mean Square)算法的有源噪声控制方法,考虑到相位信息在有源噪声控制(Active Noise Control,ANC)中的重要性,DCRN...本文提出一种利用双解码卷积循环网络(Dual-decoder Convolutional Recurrent Network,DCRN)代替FxLMS(Filtered-x Least Mean Square)算法的有源噪声控制方法,考虑到相位信息在有源噪声控制(Active Noise Control,ANC)中的重要性,DCRN网络的输入特征为噪声信号的复数频谱(包括实部谱和虚部谱).网络结构中,采用编码模块从噪声复数频谱中提取特征,利用双解码模块分别估计网络输出的实部谱和虚部谱,采用参数共享机制和组策略以降低训练参数的数量并提高网络的学习能力和泛化能力.特别是针对风噪声,选用新的损失函数以及对训练数据进行正则化处理以提升DCRN的性能.实验结果表明,DCRN方法在仿真环境与有源降噪耳机环境下对一般噪声和风噪声都表现出良好的降噪性能和鲁棒性.展开更多
【目的】为解决群养环境下生猪音频难以分离与识别的问题,提出基于欠定盲源分离与E C A-EfficientNetV2的生猪状态音频识别方法。【方法】以仿真群养环境下4类生猪音频信号作为观测信号,将信号稀疏表示后,通过层次聚类估计出信号混合矩...【目的】为解决群养环境下生猪音频难以分离与识别的问题,提出基于欠定盲源分离与E C A-EfficientNetV2的生猪状态音频识别方法。【方法】以仿真群养环境下4类生猪音频信号作为观测信号,将信号稀疏表示后,通过层次聚类估计出信号混合矩阵,并利用lp范数重构算法求解lp范数最小值以完成生猪音频信号重构。将重构信号转化为声谱图,分为进食声、咆哮声、哼叫声和发情声4类,利用ECA-EfficientNetV2网络模型识别音频,获取生猪状态。【结果】混合矩阵估计的归一化均方误差最低为3.266×10^(−4),分离重构的音频信噪比在3.254~4.267 dB之间。声谱图经ECA-EfficientNetV2识别检测,准确率高达98.35%;与经典卷积神经网络ResNet50和VGG16对比,准确率分别提升2.88和1.81个百分点;与原EfficientNetV2相比,准确率降低0.52个百分点,但模型参数量减少33.56%,浮点运算量(FLOPs)降低1.86 G,推理时间减少9.40 ms。【结论】基于盲源分离及改进EfficientNetV2的方法,轻量且高效地实现了分离与识别群养生猪音频信号。展开更多
文摘Cases are presented to reveal how modern computerised infra-red thermal imaging has the potential to assist in early breast cancer detection. The history of thermography and some recent controversies surrounding mammography are discussed. Examples of thermal imaging combined with naturopathic interventions are described. Since 2002, more than 8000 women in New Zealand have chosen to include thermal imaging as a part of their breast health management. Breast thermal imaging combined with relevant health advice, resulted in a perceived worthwhile benefit to patients in managing overall health.
文摘A new method to detect multiple outliers in multivariate data is proposed. It is a combination of minimum subsets, resampling and self-organizing map (SOM) algorithm introduced by Kohonen,which provides a robust way with neural network. In this method, the number and organization of the neurons are selected by the characteristics of the spectra, e.g., the spectra data are often changed linearly with the concentration of the components and are often measured repeatedly, etc. So the spatial distribution of the neurons can be arranged by this characteristic. With this method, all the outliers in the spectra can be detected, which cannot be solved by the traditional method, and the speed of computation is higher than that of the traditional neural network method. The results of the simulation and the experiment show that this method is simple, effective, intuitionistic and all the outliers in the spectra can be detected in a short time. It is useful when associated with the regression model in the near infra-red research.
基金supported by grants awarded by the National Major Scientic Research Program of China(Grant No.2011CB910404)the National Nature Science Foundation of China(Grant No.61227017)+1 种基金the State Key Basic Research Development Program of China(2012CB518103)National Outstanding Young Scientist Award of China(61425006).
文摘Alzheimers disease(AD)is a chronic neurodegenerative disease.The symptoms include memoryand spatial learning dificulties,language disorders,and loss of motivation,which get worse overtime,eventually ending in death.No ffective treatments are available for AD,currently.Currenttreatments only attenuate symptoms temporarily and are associated with severe side ffects.Nearinfra-red(NIR)light has been studied for a long time.We investigated the effect of NIR on ADusing a transgenic mouse model,which was obtained by co-injecting two vectors carrying ADmutations in amyloid precursor protein(APP)and presenilin-i(PSEN1)into C57BL/6J mice.The irradiation equipment consisted of an accommodating box and an LED array.The wave-length of NIR light emitted from LED was between 1040 nm and 1090 nm.The power densitydelivered at the level of the mice was approximately 15 mW/cm^(2),Firstly,we treated the micewith NIR for 40 days,Then,the irradiation was suspended for 28 days.Finally,another 15 daystreatment was brought to mice.We conducted Morris water maze and immunofluorescenceanalysis to evaluate the effects of treatment.Immunofuorescence analysis was based on mea-suring the quantity of plaques in mouse brain slices,Our results show that NIR light improvesmemory and spatial learning ability and reduces plaques moderately.NIR light represents apotential treatment for AD.
文摘In-process damage to a cutting tool degrades the surfacenish of the job shaped by machining and causes a signicantnancial loss.This stimulates the need for Tool Condition Monitoring(TCM)to assist detection of failure before it extends to the worse phase.Machine Learning(ML)based TCM has been extensively explored in the last decade.However,most of the research is now directed toward Deep Learning(DL).The“Deep”formulation,hierarchical compositionality,distributed representation and end-to-end learning of Neural Nets need to be explored to create a generalized TCM framework to perform eciently in a high-noise environment of cross-domain machining.With this motivation,the design of dierent CNN(Convolutional Neural Network)architectures such as AlexNet,ResNet-50,LeNet-5,and VGG-16 is presented in this paper.Real-time spindle vibrations corresponding to healthy and various faulty congurations of milling cutter were acquired.This data was transformed into the time-frequency domain and further processed by proposed architectures in graphical form,i.e.,spectrogram.The model is trained,tested,and validated considering dierent datasets and showcased promising results.
文摘Introduction: Infra-red (IR) thermometry is a safe and valid method to determine internal and surface temperature in human subjects. Under conditions of brain damage (head injury or stroke) knowledge of changes in the temperature of intracranial tissue is justified because of the vulnerability of neurons to accelerated damage at temperatures at the upper end of the febrile range. Aim: To determine the temperature at the inner canthus (IC) of the eye as a potential surrogate for brain temperature. Methods: Invasive monitoring of deep brain structures, lateral ventricle and deep white matter. IR temperature readings obtained at right and left IC. Results: ?Strong correlations were evident between R and L IC and brain. Close, as well as poor, agreement between?? sites was shown in some patients and at some times. For right hemispheric lesions four had a better correlation between TbrV and TRIC when compared to TLIC.? When the correlation between TbrV and TLIC was better compared to TbrV and TRIC, four had a predominant right hemispheric lesion. Conclusions: Improved techniques for IR thermal imaging accuracy at the bedside has the potential to improve temperature measurement agreement. The predominant lesion side may have a bearing on maximum ipsilateral IC temperature Further studies are ongoing in this pilot study population.
文摘噪声环境下语音检测准确率偏低是短波通话面临的公开挑战。当前已有方法应用有限,其根源在于难以可靠地在噪音环境下提取准确且高效的语音特征。针对上述问题,提出了一个面向短波通信的低秩方向梯度直方图(Low-rank Histogram of Oriented Gradient,LHOG)话音检测方法。首先,对目标音频源数据进行预处理,实现噪声环境下语音信息的可视化表征;然后,在HOG特征提取器中嵌入低秩化结构,缓解特征中的冗余信息,并降低噪声干扰,从而获得准确且高效的特征;最后,通过常用的SVM分类模型便可在噪声环境中准确快速地区分话音和噪声。测试结果表明,该方法的准确率达到了95.12%,误报率仅为0.96%,漏报率为13.14%。与现有主流方法的对比实验证明,该方法话音检测准确率高,资源占用少,能够有效提高短波通信侦控效率。
文摘本文提出一种利用双解码卷积循环网络(Dual-decoder Convolutional Recurrent Network,DCRN)代替FxLMS(Filtered-x Least Mean Square)算法的有源噪声控制方法,考虑到相位信息在有源噪声控制(Active Noise Control,ANC)中的重要性,DCRN网络的输入特征为噪声信号的复数频谱(包括实部谱和虚部谱).网络结构中,采用编码模块从噪声复数频谱中提取特征,利用双解码模块分别估计网络输出的实部谱和虚部谱,采用参数共享机制和组策略以降低训练参数的数量并提高网络的学习能力和泛化能力.特别是针对风噪声,选用新的损失函数以及对训练数据进行正则化处理以提升DCRN的性能.实验结果表明,DCRN方法在仿真环境与有源降噪耳机环境下对一般噪声和风噪声都表现出良好的降噪性能和鲁棒性.
