A heart attack disrupts the normal flow of blood to the heart muscle,potentially causing severe damage or death if not treated promptly.It can lead to long-term health complications,reduce quality of life,and signific...A heart attack disrupts the normal flow of blood to the heart muscle,potentially causing severe damage or death if not treated promptly.It can lead to long-term health complications,reduce quality of life,and significantly impact daily activities and overall well-being.Despite the growing popularity of deep learning,several drawbacks persist,such as complexity and the limitation of single-model learning.In this paper,we introduce a residual learning-based feature fusion technique to achieve high accuracy in differentiating abnormal cardiac rhythms heart sound.Combining MobileNet with DenseNet201 for feature fusion leverages MobileNet lightweight,efficient architecture with DenseNet201,dense connections,resulting in enhanced feature extraction and improved model performance with reduced computational cost.To further enhance the fusion,we employed residual learning to optimize the hierarchical features of heart abnormal sounds during training.The experimental results demonstrate that the proposed fusion method achieved an accuracy of 95.67%on the benchmark PhysioNet-2016 Spectrogram dataset.To further validate the performance,we applied it to the BreakHis dataset with a magnification level of 100X.The results indicate that the model maintains robust performance on the second dataset,achieving an accuracy of 96.55%.it highlights its consistent performance,making it a suitable for various applications.展开更多
A new lighting and enlargement on phase spectrogram （PS） and frequency spectrogram （FS） is presented in this paper. These representations result from the coupling of power spectrogram and short time Fourier transf...A new lighting and enlargement on phase spectrogram （PS） and frequency spectrogram （FS） is presented in this paper. These representations result from the coupling of power spectrogram and short time Fourier transform （STFT）. The main contribution is the construction of the 3D phase spectrogram （3DPS） and the 3D frequency spectrogram （3DFS）. These new tools allow such specific test signals as small slope linear chirp, phase jump case of musical signal analysis is reported. The main objective is to and small frequency jump to be analyzed. An application detect small frequency and phase variations in order to characterize each type of sound attack without losing the amplitude information given by power spectrogram展开更多
Acoustic array sensor device for partial discharge detection is widely used in power equipment inspection with the advantages of non-contact and precise positioning compared with partial discharge detection methods su...Acoustic array sensor device for partial discharge detection is widely used in power equipment inspection with the advantages of non-contact and precise positioning compared with partial discharge detection methods such as ultrasonic method and pulse current method.However,due to the sensitivity of the acoustic array sensor and the influence of the equipment operation site interference,the acoustic array sensor device for partial discharge type diagnosis by phase resolved partial discharge(PRPD)map might occasionally presents incorrect results,thus affecting the power equipment operation and maintenance strategy.The acoustic array sensor detection device for power equipment developed in this paper applies the array design model of equal-area multi-arm spiral with machine learning fast fourier transform clean(FFT-CLEAN)sound source localization identification algorithm to avoid the interference factors in the noise acquisition system using a single microphone and conventional beam forming algorithm,improves the spatial resolution of the acoustic array sensor device,and proposes an acoustic array sensor device based on the acoustic spectrogram.The analysis and diagnosis method of discharge type of acoustic array sensor device can effectively reduce the system misjudgment caused by factors such as the resolution of the acoustic imaging device and the time domain pulse of the digital signal,and reduce the false alarm rate of the acoustic array sensor device.The proposed method is tested by selecting power cables as the object,and its effectiveness is proved by laboratory verification and field verification.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
Recently,user recognitionmethods to authenticate personal identity has attracted significant attention especially with increased availability of various internet of things(IoT)services through fifth-generation technol...Recently,user recognitionmethods to authenticate personal identity has attracted significant attention especially with increased availability of various internet of things(IoT)services through fifth-generation technology(5G)based mobile devices.The EMG signals generated inside the body with unique individual characteristics are being studied as a part of nextgeneration user recognition methods.However,there is a limitation when applying EMG signals to user recognition systems as the same operation needs to be repeated while maintaining a constant strength of muscle over time.Hence,it is necessary to conduct research on multidimensional feature transformation that includes changes in frequency features over time.In this paper,we propose a user recognition system that applies EMG signals to the short-time fourier transform(STFT),and converts the signals into EMG spectrogram images while adjusting the time-frequency resolution to extract multidimensional features.The proposed system is composed of a data pre-processing and normalization process,spectrogram image conversion process,and final classification process.The experimental results revealed that the proposed EMG spectrogram image-based user recognition system has a 95.4%accuracy performance,which is 13%higher than the EMGsignal-based system.Such a user recognition accuracy improvement was achieved by using multidimensional features,in the time-frequency domain.展开更多
Sudden wildfires cause significant global ecological damage.While satellite imagery has advanced early fire detection and mitigation,image-based systems face limitations including high false alarm rates,visual obstruc...Sudden wildfires cause significant global ecological damage.While satellite imagery has advanced early fire detection and mitigation,image-based systems face limitations including high false alarm rates,visual obstructions,and substantial computational demands,especially in complex forest terrains.To address these challenges,this study proposes a novel forest fire detection model utilizing audio classification and machine learning.We developed an audio-based pipeline using real-world environmental sound recordings.Sounds were converted into Mel-spectrograms and classified via a Convolutional Neural Network(CNN),enabling the capture of distinctive fire acoustic signatures(e.g.,crackling,roaring)that are minimally impacted by visual or weather conditions.Internet of Things(IoT)sound sensors were crucial for generating complex environmental parameters to optimize feature extraction.The CNN model achieved high performance in stratified 5-fold cross-validation(92.4%±1.6 accuracy,91.2%±1.8 F1-score)and on test data(94.93%accuracy,93.04%F1-score),with 98.44%precision and 88.32%recall,demonstrating reliability across environmental conditions.These results indicate that the audio-based approach not only improves detection reliability but also markedly reduces computational overhead compared to traditional image-based methods.The findings suggest that acoustic sensing integrated with machine learning offers a powerful,low-cost,and efficient solution for real-time forest fire monitoring in complex,dynamic environments.展开更多
Ultrasonic guided waves (GWs) can be used to evaluate long bones effectively because of the ability to provide the information of the whole bone. In this study, a joint spectrogram segmentation and ridge-extraction (J...Ultrasonic guided waves (GWs) can be used to evaluate long bones effectively because of the ability to provide the information of the whole bone. In this study, a joint spectrogram segmentation and ridge-extraction (JSSRE) method was proposed to separate multiple modes in long bones. First, the Gabor time-frequency transform was applied to obtain the spectrogram of multimodal signals. Then, a multi-class image segmentation algorithm was used to find the corresponding region of each mode in the spectrogram, including an improved watershed transform and a region growing procedure. Finally, the ridges were extracted and the time domain signals representing individual modes were reconstructed from these ridges in each region. The validations of this method were discussed by simulated multimodal signals with different signal-to-noise ratios (SNR). The correlation coefficients between the original signals without noise and the reconstructed signals were calculated to analyze the results quantitatively. The results showed that the extracted ridges were in good agreement with generated theoretical dispersion curves, and the reconstructed signals were highly related to the original signals, even under the SNR=3 dB situation.展开更多
In order to eliminate the subjectivity of wheeze diagnosis and improve the accuracy of objective detecting methods,this paper introduces a wheeze detecting method based on spectrogram entropy analysis.This algorithm m...In order to eliminate the subjectivity of wheeze diagnosis and improve the accuracy of objective detecting methods,this paper introduces a wheeze detecting method based on spectrogram entropy analysis.This algorithm mainly comprises three steps which are preprocessing,features extracting and wheeze detecting based on support vector machine(SVM).Herein,the preprocessing consists of the short-time Fourier transform(STFT) decomposition and detrending.The features are extracted from the entropy of spectrograms.The step of detrending makes the difference of the features between wheeze and normal lung sounds more obvious.Moreover,compared with the method whose decision is based on the empirical threshold,there is no uncertain detecting result any more.Results of two testing experiments show that the detecting accuracy(AC) are 97.1%and 95.7%,respectively,which proves that the proposed method could be an efficient way to detect wheeze.展开更多
The Perception Spectrogram Structure Boundary（PSSB）parameter is proposed for speech endpoint detection as a preprocess of speech or speaker recognition.At first a hearing perception speech enhancement is carried out...The Perception Spectrogram Structure Boundary（PSSB）parameter is proposed for speech endpoint detection as a preprocess of speech or speaker recognition.At first a hearing perception speech enhancement is carried out.Then the two-dimensional enhancement is performed upon the sound spectrogram according to the difference between the determinacy distribution characteristic of speech and the random distribution characteristic of noise.Finally a decision for endpoint was made by the PSSB parameter.Experimental results show that,in a low SNR environment from-10 dB to 10 dB,the algorithm proposed in this paper may achieve higher accuracy than the extant endpoint detection algorithms.The detection accuracy of 75.2%can be reached even in the extremely low SNR at-10 dB.Therefore it is suitable for speech endpoint detection in low-SNRs environment.展开更多
The features of an attosecond extreme ultraviolet （XUV） field are encoded in the attosecond XUV spectrogram. We investigate the effect of the temporal structures of attosecond XUV fields on the attosecond streaking ...The features of an attosecond extreme ultraviolet （XUV） field are encoded in the attosecond XUV spectrogram. We investigate the effect of the temporal structures of attosecond XUV fields on the attosecond streaking spectrogram. Factors such as the number of attosecond XUV pulses and the temporal chirp of attosecond XUV pulses are considered. Results indicate that unlike the attosecond streaking spectrogram for an attosecond XUV field with two pulses of a half-cycle separation of streaking field, the spectrogram for the attosecond XUV field with three pulses demonstrates fine spectral fringes in separated traces.展开更多
A method of drawing color spectrogram of speech by using microcomputer is described in this paper , and referred to the metod of drawing spectrogram by computer . With the software and no addition any other aqripment....A method of drawing color spectrogram of speech by using microcomputer is described in this paper , and referred to the metod of drawing spectrogram by computer . With the software and no addition any other aqripment., we can draw color three - dimension spectrogram ( or black -white spectrogram without color monitor ), and it is similar to spectrogram of sonagrapher .展开更多
Cardiovascular diseases(CVDs)remain one of the foremost causes of death globally;hence,the need for several must-have,advanced automated diagnostic solutions towards early detection and intervention.Traditional auscul...Cardiovascular diseases(CVDs)remain one of the foremost causes of death globally;hence,the need for several must-have,advanced automated diagnostic solutions towards early detection and intervention.Traditional auscultation of cardiovascular sounds is heavily reliant on clinical expertise and subject to high variability.To counter this limitation,this study proposes an AI-driven classification system for cardiovascular sounds whereby deep learning techniques are engaged to automate the detection of an abnormal heartbeat.We employ FastAI vision-learner-based convolutional neural networks(CNNs)that include ResNet,DenseNet,VGG,ConvNeXt,SqueezeNet,and AlexNet to classify heart sound recordings.Instead of raw waveform analysis,the proposed approach transforms preprocessed cardiovascular audio signals into spectrograms,which are suited for capturing temporal and frequency-wise patterns.The models are trained on the PASCAL Cardiovascular Challenge dataset while taking into consideration the recording variations,noise levels,and acoustic distortions.To demonstrate generalization,external validation using Google’s Audio set Heartbeat Sound dataset was performed using a dataset rich in cardiovascular sounds.Comparative analysis revealed that DenseNet-201,ConvNext Large,and ResNet-152 could deliver superior performance to the other architectures,achieving an accuracy of 81.50%,a precision of 85.50%,and an F1-score of 84.50%.In the process,we performed statistical significance testing,such as the Wilcoxon signed-rank test,to validate performance improvements over traditional classification methods.Beyond the technical contributions,the research underscores clinical integration,outlining a pathway in which the proposed system can augment conventional electronic stethoscopes and telemedicine platforms in the AI-assisted diagnostic workflows.We also discuss in detail issues of computational efficiency,model interpretability,and ethical considerations,particularly concerning algorithmic bias stemming from imbalanced datasets and the need for real-time processing in clinical settings.The study describes a scalable,automated system combining deep learning,feature extraction using spectrograms,and external validation that can assist healthcare providers in the early and accurate detection of cardiovascular disease.AI-driven solutions can be viable in improving access,reducing delays in diagnosis,and ultimately even the continued global burden of heart disease.展开更多
Cardiovascular diseases(CVDs)are the leading cause of mortality worldwide,necessitating efficient diagnostic tools.This study develops and validates a deep learning framework for phonocardiogram(PCG)classification,foc...Cardiovascular diseases(CVDs)are the leading cause of mortality worldwide,necessitating efficient diagnostic tools.This study develops and validates a deep learning framework for phonocardiogram(PCG)classification,focusing on model generalizability and robustness.Initially,a ResNet-18 model was trained on the PhysioNet 2016 dataset,achieving high accuracy.To assess real-world viability,we conducted extensive external validation on the HLS-CMDS dataset.We performed four key experiments:(1)Fine-tuning the PhysioNet-trained model for binary(Normal/Abnormal)classification on HLS-CMDS,achieving 88%accuracy.(2)Fine-tuning the same model for multiclass classification(Normal,Murmur,Extra Sound,Rhythm Disorder),which yielded 86%accuracy.(3)Retraining a ResNet-18 model with ImageNet weights directly on the HLS-CMDS data,which improved multi-class accuracy to 89%,demonstrating the benefit of domain-specific feature learning on the target dataset.(4)A novel stress test evaluating the retrained model on computationally separated heart sounds from mixed heart-lung recordings,which revealed a significant performance drop to 41%accuracy.This highlights the model’s sensitivity to signal processing artifacts.Our findings underscore the importance of external validation and demonstrate that while deep learning models can generalize across datasets,their performance is heavily influenced by training strategy and their robustness to preprocessing artifacts remains a critical challenge for clinical deployment.展开更多
To overcome the challenges of poor real-time performance,limited scalability,and low intelligence in conventional jamming pattern recognition methods,this paper proposes a method based on Wavelet Packet Decomposition(...To overcome the challenges of poor real-time performance,limited scalability,and low intelligence in conventional jamming pattern recognition methods,this paper proposes a method based on Wavelet Packet Decomposition(WPD)and enhanced deep learning techniques.In the proposed method,an agent at the receiver processes the received signal using WPD to generate an initial Spectrogram Waterfall(SW),which is subsequently segmented using a sliding window to serve as the input for the jamming recognition network.The network employs a bilateral filter to preprocess the input SW,thereby enhancing the edge features of the jamming signals.To extract abstract features,depthwise separable convolution is utilized instead of traditional convolution,thereby reducing the network’s parameter count and enhancing real-time performance.A pyramid pooling layer is integrated before the fully connected layer to enable the network to process input SW of varying sizes,thus enhancing scalability.During network training,adaptive moment estimation is employed as the optimizer,allowing the network to dynamically adjust the learning rate and accelerate convergence.A comprehensive comparison between the proposed jamming recognition network and six other models is conducted,along with Ablation Experiments(AE)based on numerical simulations.Simulation results demonstrate that the proposed method based on WPD and enhanced deep learning achieves high-precision recognition of various jamming patterns while maintaining a favorable balance among prediction accuracy,network complexity,and prediction time.展开更多
文摘A heart attack disrupts the normal flow of blood to the heart muscle,potentially causing severe damage or death if not treated promptly.It can lead to long-term health complications,reduce quality of life,and significantly impact daily activities and overall well-being.Despite the growing popularity of deep learning,several drawbacks persist,such as complexity and the limitation of single-model learning.In this paper,we introduce a residual learning-based feature fusion technique to achieve high accuracy in differentiating abnormal cardiac rhythms heart sound.Combining MobileNet with DenseNet201 for feature fusion leverages MobileNet lightweight,efficient architecture with DenseNet201,dense connections,resulting in enhanced feature extraction and improved model performance with reduced computational cost.To further enhance the fusion,we employed residual learning to optimize the hierarchical features of heart abnormal sounds during training.The experimental results demonstrate that the proposed fusion method achieved an accuracy of 95.67%on the benchmark PhysioNet-2016 Spectrogram dataset.To further validate the performance,we applied it to the BreakHis dataset with a magnification level of 100X.The results indicate that the model maintains robust performance on the second dataset,achieving an accuracy of 96.55%.it highlights its consistent performance,making it a suitable for various applications.
文摘A new lighting and enlargement on phase spectrogram （PS） and frequency spectrogram （FS） is presented in this paper. These representations result from the coupling of power spectrogram and short time Fourier transform （STFT）. The main contribution is the construction of the 3D phase spectrogram （3DPS） and the 3D frequency spectrogram （3DFS）. These new tools allow such specific test signals as small slope linear chirp, phase jump case of musical signal analysis is reported. The main objective is to and small frequency jump to be analyzed. An application detect small frequency and phase variations in order to characterize each type of sound attack without losing the amplitude information given by power spectrogram
基金This work was supported by the science and technology project of State Grid Shanghai Municipal Electric Power Company(No.52090020007F)National Key R&D Program of China(2017YFB0902800).
文摘Acoustic array sensor device for partial discharge detection is widely used in power equipment inspection with the advantages of non-contact and precise positioning compared with partial discharge detection methods such as ultrasonic method and pulse current method.However,due to the sensitivity of the acoustic array sensor and the influence of the equipment operation site interference,the acoustic array sensor device for partial discharge type diagnosis by phase resolved partial discharge(PRPD)map might occasionally presents incorrect results,thus affecting the power equipment operation and maintenance strategy.The acoustic array sensor detection device for power equipment developed in this paper applies the array design model of equal-area multi-arm spiral with machine learning fast fourier transform clean(FFT-CLEAN)sound source localization identification algorithm to avoid the interference factors in the noise acquisition system using a single microphone and conventional beam forming algorithm,improves the spatial resolution of the acoustic array sensor device,and proposes an acoustic array sensor device based on the acoustic spectrogram.The analysis and diagnosis method of discharge type of acoustic array sensor device can effectively reduce the system misjudgment caused by factors such as the resolution of the acoustic imaging device and the time domain pulse of the digital signal,and reduce the false alarm rate of the acoustic array sensor device.The proposed method is tested by selecting power cables as the object,and its effectiveness is proved by laboratory verification and field verification.
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.
基金supported by Basic Science Research Program through the National Research Foundation of Korea(NRF)funded by the Ministry of Education(No.2017R1A6A1A03015496)the National Research Foundation of Korea(NRF)grant funded by the Korea government(MSIT)(No.NRF-2021R1A2C1014033).
文摘Recently,user recognitionmethods to authenticate personal identity has attracted significant attention especially with increased availability of various internet of things(IoT)services through fifth-generation technology(5G)based mobile devices.The EMG signals generated inside the body with unique individual characteristics are being studied as a part of nextgeneration user recognition methods.However,there is a limitation when applying EMG signals to user recognition systems as the same operation needs to be repeated while maintaining a constant strength of muscle over time.Hence,it is necessary to conduct research on multidimensional feature transformation that includes changes in frequency features over time.In this paper,we propose a user recognition system that applies EMG signals to the short-time fourier transform(STFT),and converts the signals into EMG spectrogram images while adjusting the time-frequency resolution to extract multidimensional features.The proposed system is composed of a data pre-processing and normalization process,spectrogram image conversion process,and final classification process.The experimental results revealed that the proposed EMG spectrogram image-based user recognition system has a 95.4%accuracy performance,which is 13%higher than the EMGsignal-based system.Such a user recognition accuracy improvement was achieved by using multidimensional features,in the time-frequency domain.
基金funded by the Directorate of Research and Community Service,Directorate General of Research and Development,Ministry of Higher Education,Science and Technologyin accordance with the Implementation Contract for the Operational Assistance Program for State Universities,Research Program Number:109/C3/DT.05.00/PL/2025.
文摘Sudden wildfires cause significant global ecological damage.While satellite imagery has advanced early fire detection and mitigation,image-based systems face limitations including high false alarm rates,visual obstructions,and substantial computational demands,especially in complex forest terrains.To address these challenges,this study proposes a novel forest fire detection model utilizing audio classification and machine learning.We developed an audio-based pipeline using real-world environmental sound recordings.Sounds were converted into Mel-spectrograms and classified via a Convolutional Neural Network(CNN),enabling the capture of distinctive fire acoustic signatures(e.g.,crackling,roaring)that are minimally impacted by visual or weather conditions.Internet of Things(IoT)sound sensors were crucial for generating complex environmental parameters to optimize feature extraction.The CNN model achieved high performance in stratified 5-fold cross-validation(92.4%±1.6 accuracy,91.2%±1.8 F1-score)and on test data(94.93%accuracy,93.04%F1-score),with 98.44%precision and 88.32%recall,demonstrating reliability across environmental conditions.These results indicate that the audio-based approach not only improves detection reliability but also markedly reduces computational overhead compared to traditional image-based methods.The findings suggest that acoustic sensing integrated with machine learning offers a powerful,low-cost,and efficient solution for real-time forest fire monitoring in complex,dynamic environments.
基金supported by the National Natural Science Foundation of China(Grant No. 11174060)the PhD Programs Foundation of the Ministry of Education of China(Grant Nos. 20090071110066 and 20110071130004)the New Century Excellent Talents of the Ministry of Education of China(Grant No. NCET-10-0349)
文摘Ultrasonic guided waves (GWs) can be used to evaluate long bones effectively because of the ability to provide the information of the whole bone. In this study, a joint spectrogram segmentation and ridge-extraction (JSSRE) method was proposed to separate multiple modes in long bones. First, the Gabor time-frequency transform was applied to obtain the spectrogram of multimodal signals. Then, a multi-class image segmentation algorithm was used to find the corresponding region of each mode in the spectrogram, including an improved watershed transform and a region growing procedure. Finally, the ridges were extracted and the time domain signals representing individual modes were reconstructed from these ridges in each region. The validations of this method were discussed by simulated multimodal signals with different signal-to-noise ratios (SNR). The correlation coefficients between the original signals without noise and the reconstructed signals were calculated to analyze the results quantitatively. The results showed that the extracted ridges were in good agreement with generated theoretical dispersion curves, and the reconstructed signals were highly related to the original signals, even under the SNR=3 dB situation.
文摘In order to eliminate the subjectivity of wheeze diagnosis and improve the accuracy of objective detecting methods,this paper introduces a wheeze detecting method based on spectrogram entropy analysis.This algorithm mainly comprises three steps which are preprocessing,features extracting and wheeze detecting based on support vector machine(SVM).Herein,the preprocessing consists of the short-time Fourier transform(STFT) decomposition and detrending.The features are extracted from the entropy of spectrograms.The step of detrending makes the difference of the features between wheeze and normal lung sounds more obvious.Moreover,compared with the method whose decision is based on the empirical threshold,there is no uncertain detecting result any more.Results of two testing experiments show that the detecting accuracy(AC) are 97.1%and 95.7%,respectively,which proves that the proposed method could be an efficient way to detect wheeze.
基金supported by the National Natural Science Foundation of China.(61071215,61271359,61372146)
文摘The Perception Spectrogram Structure Boundary（PSSB）parameter is proposed for speech endpoint detection as a preprocess of speech or speaker recognition.At first a hearing perception speech enhancement is carried out.Then the two-dimensional enhancement is performed upon the sound spectrogram according to the difference between the determinacy distribution characteristic of speech and the random distribution characteristic of noise.Finally a decision for endpoint was made by the PSSB parameter.Experimental results show that,in a low SNR environment from-10 dB to 10 dB,the algorithm proposed in this paper may achieve higher accuracy than the extant endpoint detection algorithms.The detection accuracy of 75.2%can be reached even in the extremely low SNR at-10 dB.Therefore it is suitable for speech endpoint detection in low-SNRs environment.
基金supported in part by the BK21 project, the Basic Research Program (No. KRF-2008-313-C00356) funded by Korean Research Foundationthe Global Research Laboratory Program funded by National Research Foundation (No. 2009-00439).
文摘The features of an attosecond extreme ultraviolet （XUV） field are encoded in the attosecond XUV spectrogram. We investigate the effect of the temporal structures of attosecond XUV fields on the attosecond streaking spectrogram. Factors such as the number of attosecond XUV pulses and the temporal chirp of attosecond XUV pulses are considered. Results indicate that unlike the attosecond streaking spectrogram for an attosecond XUV field with two pulses of a half-cycle separation of streaking field, the spectrogram for the attosecond XUV field with three pulses demonstrates fine spectral fringes in separated traces.
文摘A method of drawing color spectrogram of speech by using microcomputer is described in this paper , and referred to the metod of drawing spectrogram by computer . With the software and no addition any other aqripment., we can draw color three - dimension spectrogram ( or black -white spectrogram without color monitor ), and it is similar to spectrogram of sonagrapher .
基金funded by the deanship of scientific research(DSR),King Abdulaziz University,Jeddah,under grant No.(G-1436-611-309).
文摘Cardiovascular diseases(CVDs)remain one of the foremost causes of death globally;hence,the need for several must-have,advanced automated diagnostic solutions towards early detection and intervention.Traditional auscultation of cardiovascular sounds is heavily reliant on clinical expertise and subject to high variability.To counter this limitation,this study proposes an AI-driven classification system for cardiovascular sounds whereby deep learning techniques are engaged to automate the detection of an abnormal heartbeat.We employ FastAI vision-learner-based convolutional neural networks(CNNs)that include ResNet,DenseNet,VGG,ConvNeXt,SqueezeNet,and AlexNet to classify heart sound recordings.Instead of raw waveform analysis,the proposed approach transforms preprocessed cardiovascular audio signals into spectrograms,which are suited for capturing temporal and frequency-wise patterns.The models are trained on the PASCAL Cardiovascular Challenge dataset while taking into consideration the recording variations,noise levels,and acoustic distortions.To demonstrate generalization,external validation using Google’s Audio set Heartbeat Sound dataset was performed using a dataset rich in cardiovascular sounds.Comparative analysis revealed that DenseNet-201,ConvNext Large,and ResNet-152 could deliver superior performance to the other architectures,achieving an accuracy of 81.50%,a precision of 85.50%,and an F1-score of 84.50%.In the process,we performed statistical significance testing,such as the Wilcoxon signed-rank test,to validate performance improvements over traditional classification methods.Beyond the technical contributions,the research underscores clinical integration,outlining a pathway in which the proposed system can augment conventional electronic stethoscopes and telemedicine platforms in the AI-assisted diagnostic workflows.We also discuss in detail issues of computational efficiency,model interpretability,and ethical considerations,particularly concerning algorithmic bias stemming from imbalanced datasets and the need for real-time processing in clinical settings.The study describes a scalable,automated system combining deep learning,feature extraction using spectrograms,and external validation that can assist healthcare providers in the early and accurate detection of cardiovascular disease.AI-driven solutions can be viable in improving access,reducing delays in diagnosis,and ultimately even the continued global burden of heart disease.
基金Deanship of Graduate Studies and Scientific Research at Qassim University through grant number(QU-APC-2025)。
文摘Cardiovascular diseases(CVDs)are the leading cause of mortality worldwide,necessitating efficient diagnostic tools.This study develops and validates a deep learning framework for phonocardiogram(PCG)classification,focusing on model generalizability and robustness.Initially,a ResNet-18 model was trained on the PhysioNet 2016 dataset,achieving high accuracy.To assess real-world viability,we conducted extensive external validation on the HLS-CMDS dataset.We performed four key experiments:(1)Fine-tuning the PhysioNet-trained model for binary(Normal/Abnormal)classification on HLS-CMDS,achieving 88%accuracy.(2)Fine-tuning the same model for multiclass classification(Normal,Murmur,Extra Sound,Rhythm Disorder),which yielded 86%accuracy.(3)Retraining a ResNet-18 model with ImageNet weights directly on the HLS-CMDS data,which improved multi-class accuracy to 89%,demonstrating the benefit of domain-specific feature learning on the target dataset.(4)A novel stress test evaluating the retrained model on computationally separated heart sounds from mixed heart-lung recordings,which revealed a significant performance drop to 41%accuracy.This highlights the model’s sensitivity to signal processing artifacts.Our findings underscore the importance of external validation and demonstrate that while deep learning models can generalize across datasets,their performance is heavily influenced by training strategy and their robustness to preprocessing artifacts remains a critical challenge for clinical deployment.
基金supported by National Natural Science Foundation of China under Grant U23A20279China Electronics Tian’ao Innovation Theory and Technology Group Fund under Grand 20221193-04-04.
文摘To overcome the challenges of poor real-time performance,limited scalability,and low intelligence in conventional jamming pattern recognition methods,this paper proposes a method based on Wavelet Packet Decomposition(WPD)and enhanced deep learning techniques.In the proposed method,an agent at the receiver processes the received signal using WPD to generate an initial Spectrogram Waterfall(SW),which is subsequently segmented using a sliding window to serve as the input for the jamming recognition network.The network employs a bilateral filter to preprocess the input SW,thereby enhancing the edge features of the jamming signals.To extract abstract features,depthwise separable convolution is utilized instead of traditional convolution,thereby reducing the network’s parameter count and enhancing real-time performance.A pyramid pooling layer is integrated before the fully connected layer to enable the network to process input SW of varying sizes,thus enhancing scalability.During network training,adaptive moment estimation is employed as the optimizer,allowing the network to dynamically adjust the learning rate and accelerate convergence.A comprehensive comparison between the proposed jamming recognition network and six other models is conducted,along with Ablation Experiments(AE)based on numerical simulations.Simulation results demonstrate that the proposed method based on WPD and enhanced deep learning achieves high-precision recognition of various jamming patterns while maintaining a favorable balance among prediction accuracy,network complexity,and prediction time.
