Mazen Megahed, U. Jain, Michael T. Leasure, Adam A. Butchy
There are multiple modalities used to diagnose abnormalities of the heart consisting of various invasive and noninvasive tests. Patients may undergo multiple tests, progressing to more invasive methods at the expense of patient risk and cost to the pair. HEARTio, through machine learning and algorithmic processing our proprietary software, hopes to improve the accuracy of the electrocardiography: a century old technology and the most commonly performed cardiac test. It is used to diagnose heart attacks, heart rhythm problems and operates as the gateway testing for patients undergoing cardiac evaluation. Myocardial infarction, or heart attacks, affect almost 800,000 Americans yearly [7] with time to treatment being the most important factor in recovery and therapy. We show in this paper that we are able to localize and detect myocardial infarctions at an accuracy above 99% by applying our system to the PTB database.
{"title":"Localization of Myocardial Infarction from 12 Lead ECG Empowered with Novel Machine Learning","authors":"Mazen Megahed, U. Jain, Michael T. Leasure, Adam A. Butchy","doi":"10.1145/3386164.3389084","DOIUrl":"https://doi.org/10.1145/3386164.3389084","url":null,"abstract":"There are multiple modalities used to diagnose abnormalities of the heart consisting of various invasive and noninvasive tests. Patients may undergo multiple tests, progressing to more invasive methods at the expense of patient risk and cost to the pair. HEARTio, through machine learning and algorithmic processing our proprietary software, hopes to improve the accuracy of the electrocardiography: a century old technology and the most commonly performed cardiac test. It is used to diagnose heart attacks, heart rhythm problems and operates as the gateway testing for patients undergoing cardiac evaluation. Myocardial infarction, or heart attacks, affect almost 800,000 Americans yearly [7] with time to treatment being the most important factor in recovery and therapy. We show in this paper that we are able to localize and detect myocardial infarctions at an accuracy above 99% by applying our system to the PTB database.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"136 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133395979","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The human gait contains the information of lower limb movement posture. Gait recognition and analysis can provide control information for lower limb rehabilitation robots. In different gait phases, the position of the foot touching the ground is different. Therefore, gait classification can be performed according to plantar pressure. This study investigated the changes in plantar pressure information during human lower limb walking. Firstly, the STM32 single-chip microcomputer is used to collect the plantar pressure information detected by the high dynamic force sensing resistor(HD-FSR); secondly, the collected information is transmitted to the upper computer through Bluetooth; finally, all kinds of gait phases pressure features are analyzed and classified. Since the two constant parameters in the Support Vector Machine(SVM) classification algorithm have a great influence on the classification effect, a SVM classification algorithm based on the Particle Swarm Optimization(PSO) is proposed to realize the recognition of different phases.
{"title":"The Research of Gait Recognition Based on High Dynamic Force Sensing Resistor","authors":"Peng Yang, Xiaodong Cai, Yanli Geng, Lingling Chen","doi":"10.1145/3386164.3386172","DOIUrl":"https://doi.org/10.1145/3386164.3386172","url":null,"abstract":"The human gait contains the information of lower limb movement posture. Gait recognition and analysis can provide control information for lower limb rehabilitation robots. In different gait phases, the position of the foot touching the ground is different. Therefore, gait classification can be performed according to plantar pressure. This study investigated the changes in plantar pressure information during human lower limb walking. Firstly, the STM32 single-chip microcomputer is used to collect the plantar pressure information detected by the high dynamic force sensing resistor(HD-FSR); secondly, the collected information is transmitted to the upper computer through Bluetooth; finally, all kinds of gait phases pressure features are analyzed and classified. Since the two constant parameters in the Support Vector Machine(SVM) classification algorithm have a great influence on the classification effect, a SVM classification algorithm based on the Particle Swarm Optimization(PSO) is proposed to realize the recognition of different phases.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"74 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124736160","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper has been discussed about speech recognition using spectrogram images and deep convolution neural network(CNN) of Uzbek spoken digits. Spectrogram images from speech signal were generated and it were used for deep CNN training. Presented CNN model contains 3 convolution layers and 2 fully connected layers that discriminative features can be divided and estimated of spectrogram images by those layers. In current research period, dataset of Uzbek spoken digits were made and in based on presented CNN model they were trained. Testing results shows that, proposed approach for Uzbek spoken digits classified 100% accuracy.
{"title":"Image Approach to Speech Recognition on CNN","authors":"M. Musaev, Ilyos Khujayorov, M. Ochilov","doi":"10.1145/3386164.3389100","DOIUrl":"https://doi.org/10.1145/3386164.3389100","url":null,"abstract":"In this paper has been discussed about speech recognition using spectrogram images and deep convolution neural network(CNN) of Uzbek spoken digits. Spectrogram images from speech signal were generated and it were used for deep CNN training. Presented CNN model contains 3 convolution layers and 2 fully connected layers that discriminative features can be divided and estimated of spectrogram images by those layers. In current research period, dataset of Uzbek spoken digits were made and in based on presented CNN model they were trained. Testing results shows that, proposed approach for Uzbek spoken digits classified 100% accuracy.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122720469","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Antoun Yaacoub, Z. Assaghir, S. Makki, Radwan Almokdad
Apathy is a behavioral and personality change and is generally defined as a loss of motivation. In this study, we will explore the detection of apathy in two phases. An apathy detection phase, and relevant features identification phase. In the first one, we apply micro-facial expressions detection systems and counter for the purpose of diagnosing clinical manifestation of apathy from a video for Lebanese citizens. The method works by applying Histogram of Oriented Gradients (HOG) as a feature descriptor on video dataset of spontaneous micro facial movements. Micro-facial expressions appear by video recording participants reacting to emotional stimulating COPE cards. Results are compared to Lille Apathy Rating Scale LARS scores. Kappa agreement was calculated to be 95.96% showing the proposed classification method has a high accuracy of estimation. In a second phase, we aim to identify the demographics and habits that might be affecting the manifestation of apathy using machine leaning algorithms. A statistical model is built based on the results to identify the characteristics that affect the manifestation of apathy by analyzing the data and making a statistical description. Using a sample of 470 participants, we base our results on the decision tree (CART) combined with logistic regression. Finally, we found that insomnia, genetic background and stress are the most important features that influence the manifestation of apathy, with an accuracy of 96.7%.
冷漠是一种行为和性格的改变，通常被定义为失去动力。在本研究中，我们将从两个阶段探讨冷漠的检测。一个冷漠检测阶段，以及相关特征识别阶段。在第一篇文章中，我们应用微面部表情检测系统和计数器来诊断黎巴嫩公民视频中冷漠的临床表现。该方法将定向梯度直方图(Histogram of Oriented Gradients, HOG)作为特征描述符应用于面部微运动视频数据集。通过视频记录参与者对情绪刺激的COPE卡片的反应，可以看到微面部表情。结果与Lille冷漠评定量表(LARS)评分进行比较。Kappa一致性为95.96%，表明所提出的分类方法具有较高的估计精度。在第二阶段，我们的目标是使用机器学习算法确定可能影响冷漠表现的人口统计数据和习惯。在此基础上建立统计模型，通过对数据的分析和统计描述，找出影响冷漠表现的特征。使用470参与者的样本，我们基于决策树(CART)结合逻辑回归的结果。最后，我们发现失眠、遗传背景和压力是影响冷漠表现的最重要特征，准确率为96.7%。
{"title":"Diagnosing Clinical Manifestation of Apathy Using Machine Learning and Micro-facial Expressions Detection","authors":"Antoun Yaacoub, Z. Assaghir, S. Makki, Radwan Almokdad","doi":"10.1145/3386164.3386174","DOIUrl":"https://doi.org/10.1145/3386164.3386174","url":null,"abstract":"Apathy is a behavioral and personality change and is generally defined as a loss of motivation. In this study, we will explore the detection of apathy in two phases. An apathy detection phase, and relevant features identification phase. In the first one, we apply micro-facial expressions detection systems and counter for the purpose of diagnosing clinical manifestation of apathy from a video for Lebanese citizens. The method works by applying Histogram of Oriented Gradients (HOG) as a feature descriptor on video dataset of spontaneous micro facial movements. Micro-facial expressions appear by video recording participants reacting to emotional stimulating COPE cards. Results are compared to Lille Apathy Rating Scale LARS scores. Kappa agreement was calculated to be 95.96% showing the proposed classification method has a high accuracy of estimation. In a second phase, we aim to identify the demographics and habits that might be affecting the manifestation of apathy using machine leaning algorithms. A statistical model is built based on the results to identify the characteristics that affect the manifestation of apathy by analyzing the data and making a statistical description. Using a sample of 470 participants, we base our results on the decision tree (CART) combined with logistic regression. Finally, we found that insomnia, genetic background and stress are the most important features that influence the manifestation of apathy, with an accuracy of 96.7%.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122057976","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
There have been many studies regarding visual simulations that consider the characteristic movement of auroras. We have proposed a method of generating animation of auroras in the desired form and in the desired location visualized by the users. This study is based on the method proposed by Kojima et. al [5], in which shape control is performed comparatively easily through parameter adjustment. With this method, an artificial 2D distributed simulation of auroras, comprised of inflow points for charged particles flowing from space, has been produced. The curtain-shaped movement of auroras can be reproduced by applying a kinetic model using an electromagnetic field calculation and a fluid calculation within the simulation space. We can see that the reproduction of aurora-specific movement is dependent on the initial value of the current volume flowing from the various flow points. In this way, we attempted to control the shape of the desired aurora by controlling the current flow. In this study, we extracted two frames from the live-captured aurora video, and, set the initial distribution and target distribution of the aurora by reproducing the respective aurora distributions in 3D. As the respective distributions feature flow limits of charged particles forming an aurora 100 km above the ground, and many aurora video images often capture the horizon, we set the camera position as the point of origin and calculated the world coordinates for the lowest section of the aurora. A genetic algorithm was used to optimize the current flows. We set the cost function as the difference between the electric potential of the target shape and the electric potential based on the simulation results for the coordinates of each flow point. In addition, the number of searched parameters were reduced, assuming that the current distribution flowing to each flow point changes along with the initial shape functionally by expanding this function in a Fourier series, General shape control made possible through optimization. In the future works, we aim to increase control accuracy and gain the ability to control complex shapes.
{"title":"Procedural Animation of Aurora and its Optimization for Keyframe Animation","authors":"Tomokazu Ishikawa, Ryota Nakazato, I. Matsuda","doi":"10.1145/3386164.3389098","DOIUrl":"https://doi.org/10.1145/3386164.3389098","url":null,"abstract":"There have been many studies regarding visual simulations that consider the characteristic movement of auroras. We have proposed a method of generating animation of auroras in the desired form and in the desired location visualized by the users. This study is based on the method proposed by Kojima et. al [5], in which shape control is performed comparatively easily through parameter adjustment. With this method, an artificial 2D distributed simulation of auroras, comprised of inflow points for charged particles flowing from space, has been produced. The curtain-shaped movement of auroras can be reproduced by applying a kinetic model using an electromagnetic field calculation and a fluid calculation within the simulation space. We can see that the reproduction of aurora-specific movement is dependent on the initial value of the current volume flowing from the various flow points. In this way, we attempted to control the shape of the desired aurora by controlling the current flow. In this study, we extracted two frames from the live-captured aurora video, and, set the initial distribution and target distribution of the aurora by reproducing the respective aurora distributions in 3D. As the respective distributions feature flow limits of charged particles forming an aurora 100 km above the ground, and many aurora video images often capture the horizon, we set the camera position as the point of origin and calculated the world coordinates for the lowest section of the aurora. A genetic algorithm was used to optimize the current flows. We set the cost function as the difference between the electric potential of the target shape and the electric potential based on the simulation results for the coordinates of each flow point. In addition, the number of searched parameters were reduced, assuming that the current distribution flowing to each flow point changes along with the initial shape functionally by expanding this function in a Fourier series, General shape control made possible through optimization. In the future works, we aim to increase control accuracy and gain the ability to control complex shapes.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123061463","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Modern embedded systems are prone to erroneous bit-flips introduced in its hardware by external disturbances such as alpha particles, electromagnetic interference or intentional external attackers. In order to protect embedded systems against these disturbances, a wide variety of software-implemented detection techniques have been proposed, a.o. by the authors of this paper. Implementing those techniques, however, can be arduous and error-prone since they have to be implemented in low-level code, e.g. assembly. To overcome this problem we propose a compiler extension, in the form of a plugin, that can automatically add any supported technique to the low-level code of the target program. We discuss the internal working of our compiler extension and conclude with a demonstration using an example program and validate the effectiveness of the introduced countermeasures by running a fault injection campaign.
{"title":"Automatic Implementation of Control Flow Error Detection Techniques","authors":"J. Vankeirsbilck, H. Hallez, J. Boydens","doi":"10.1145/3386164.3389106","DOIUrl":"https://doi.org/10.1145/3386164.3389106","url":null,"abstract":"Modern embedded systems are prone to erroneous bit-flips introduced in its hardware by external disturbances such as alpha particles, electromagnetic interference or intentional external attackers. In order to protect embedded systems against these disturbances, a wide variety of software-implemented detection techniques have been proposed, a.o. by the authors of this paper. Implementing those techniques, however, can be arduous and error-prone since they have to be implemented in low-level code, e.g. assembly. To overcome this problem we propose a compiler extension, in the form of a plugin, that can automatically add any supported technique to the low-level code of the target program. We discuss the internal working of our compiler extension and conclude with a demonstration using an example program and validate the effectiveness of the introduced countermeasures by running a fault injection campaign.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129875939","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In ascertaining the authenticity of a person's identity, the techniques that utilises Biometric Technology and its vital aspects and attributes are essential in endorsing this endeavour. Transversing the years, identification through biometric means has heightened much attention with regards to its susceptibility to assaults whilst information are being transferred. Thus, there is a dire need for the development for a safeguarding means. Steganography and watermarking methods are adopted towards the enhancement of biometric information secureness. Watermarking entails the embedding of data within file that carries the data, in safeguarding intellectual property rights and copyrights that involves music, video or image files. Meanwhile, steganography entails the concealment of data. This research puts forward an all-encompassing review pertaining the latest steganography methods that has been implemented to protect the biometric fingerprints information up-to-date. This review categorizes the fingerprint system into various sections, inclusive of methods for extracting features and matching procedure.
{"title":"All-encompassing Review of Biometric Information Protection in Fingerprints Based Steganography","authors":"M. M. Hashim, Ahmed Kamal Mohsin, M. Rahim","doi":"10.1145/3386164.3389079","DOIUrl":"https://doi.org/10.1145/3386164.3389079","url":null,"abstract":"In ascertaining the authenticity of a person's identity, the techniques that utilises Biometric Technology and its vital aspects and attributes are essential in endorsing this endeavour. Transversing the years, identification through biometric means has heightened much attention with regards to its susceptibility to assaults whilst information are being transferred. Thus, there is a dire need for the development for a safeguarding means. Steganography and watermarking methods are adopted towards the enhancement of biometric information secureness. Watermarking entails the embedding of data within file that carries the data, in safeguarding intellectual property rights and copyrights that involves music, video or image files. Meanwhile, steganography entails the concealment of data. This research puts forward an all-encompassing review pertaining the latest steganography methods that has been implemented to protect the biometric fingerprints information up-to-date. This review categorizes the fingerprint system into various sections, inclusive of methods for extracting features and matching procedure.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130116269","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The present study aims at exploiting lexicographic search(lexi-search in short) technique for modeling push broom imaging systems on board a spacecraft. Conventional techniques use a full force rigorous orbital photogrammetric model adopting iterative least squares techniques with partial derivatives of adjustment equations involving Taylor's series approximation. The proposed lexi-search technique is presented to overcome the general objection of handling rational polynomials with least square theory. The study illustrates this is excellent alternative to many bundled adjustment processes in use for camera calibration, geometric rectification, and 3D rendering.
{"title":"Lexisearch for Modeling Pushbroom Imaging Systems","authors":"S. Koduri","doi":"10.1145/3386164.3389080","DOIUrl":"https://doi.org/10.1145/3386164.3389080","url":null,"abstract":"The present study aims at exploiting lexicographic search(lexi-search in short) technique for modeling push broom imaging systems on board a spacecraft. Conventional techniques use a full force rigorous orbital photogrammetric model adopting iterative least squares techniques with partial derivatives of adjustment equations involving Taylor's series approximation. The proposed lexi-search technique is presented to overcome the general objection of handling rational polynomials with least square theory. The study illustrates this is excellent alternative to many bundled adjustment processes in use for camera calibration, geometric rectification, and 3D rendering.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120821760","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
As an important information carrier for hospital to record medical activities for patients, medical imaging report contains a large amount of technical terms and medical knowledge. In order to automatically generate computer-aided diagnosis reports, it is necessary to extract effective information from medical image reports, so as to reduce the pressure of professional physicians and better serve clinical decision-making. This paper mainly focuses on mammography medical imaging reports, analyzes the structure and contents of the reports, and deals with the imaging reports using the machine learning model, called Bi-LSTM + CRF (Bidirectional Long Short Term Memory with a Conditional Random Fields layer), in order to extract tags of the lesion, such as the position, size and shape in the imaging reports. The experimental results achieved satisfactory effort.
医学影像报告是医院记录患者医疗活动的重要信息载体，它包含了大量的专业术语和医学知识。为了自动生成计算机辅助诊断报告，有必要从医学图像报告中提取有效信息，以减轻专业医生的压力，更好地为临床决策服务。本文主要针对乳腺x线摄影医学影像报告，对报告的结构和内容进行分析，并利用Bi-LSTM + CRF (Bidirectional Long - Short Term Memory with a Conditional Random Fields layer)机器学习模型对影像报告进行处理，提取影像报告中病灶的位置、大小、形状等标签。实验结果取得了满意的效果。
{"title":"Study on Medical Imaging Reports Tagging Extraction Based on Bi-LSTM + CRF","authors":"Jiyun Li, Kaihua Li","doi":"10.1145/3386164.3389082","DOIUrl":"https://doi.org/10.1145/3386164.3389082","url":null,"abstract":"As an important information carrier for hospital to record medical activities for patients, medical imaging report contains a large amount of technical terms and medical knowledge. In order to automatically generate computer-aided diagnosis reports, it is necessary to extract effective information from medical image reports, so as to reduce the pressure of professional physicians and better serve clinical decision-making. This paper mainly focuses on mammography medical imaging reports, analyzes the structure and contents of the reports, and deals with the imaging reports using the machine learning model, called Bi-LSTM + CRF (Bidirectional Long Short Term Memory with a Conditional Random Fields layer), in order to extract tags of the lesion, such as the position, size and shape in the imaging reports. The experimental results achieved satisfactory effort.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122503827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Multi-channel blind source separation problem is a difficult but commonly met task in daily life. How to separate every target signals well using several observed signals fascinates hundreds of researchers to study on it, especially when the input channels are less than output channels. In this paper, authors analyze the characteristics of different BSS methods and propose a new method to solve under-determined multi-channel source separation problem. The proposed method has a combination structure of deep learning model and independent component analysis algorithm. Comparisons and evaluations show that using proposed method in multichannel source separation boosts separation effect to a certain degree.
{"title":"A Combination Method of Deep Learning Speech Separation and Independent Vector Analysis on Multi-channel Source Separation","authors":"Chunpeng Wang, Jie Zhu","doi":"10.1145/3386164.3386170","DOIUrl":"https://doi.org/10.1145/3386164.3386170","url":null,"abstract":"Multi-channel blind source separation problem is a difficult but commonly met task in daily life. How to separate every target signals well using several observed signals fascinates hundreds of researchers to study on it, especially when the input channels are less than output channels. In this paper, authors analyze the characteristics of different BSS methods and propose a new method to solve under-determined multi-channel source separation problem. The proposed method has a combination structure of deep learning model and independent component analysis algorithm. Comparisons and evaluations show that using proposed method in multichannel source separation boosts separation effect to a certain degree.","PeriodicalId":231209,"journal":{"name":"Proceedings of the 2019 3rd International Symposium on Computer Science and Intelligent Control","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125215428","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：