The integration of distributed energy resources (DERs) into distribution systems greatly increases the system complexity and introduces two-way power flows. Conventional protection schemes are based upon local measurements and simple linear system models, thus they cannot handle the new complexity and power flow patterns in systems with high DERs penetration. In this paper, we propose a data-driven protection framework to address the challenges induced by DERs. Considering the limited available data under fault conditions, we adopt the support vector data description (SVDD) method, a commonly used one-class classifier, for distribution system fault detection. The proposed method is tested under the IEEE 123-node test feeder and simulation results show that our proposed SVDD-based fault detection method significantly improves the robustness and resilience against DERs in comparison with conventional protection systems.
{"title":"ONE-CLASS CLASSIFIER BASED FAULT DETECTION IN DISTRIBUTION SYSTEMS WITH DISTRIBUTED ENERGY RESOURCES","authors":"Zhidi Lin, Dongliang Duan, Qi Yang, Xiang Cheng, Liuqing Yang, Shuguang Cui","doi":"10.1109/GlobalSIP.2018.8646526","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646526","url":null,"abstract":"The integration of distributed energy resources (DERs) into distribution systems greatly increases the system complexity and introduces two-way power flows. Conventional protection schemes are based upon local measurements and simple linear system models, thus they cannot handle the new complexity and power flow patterns in systems with high DERs penetration. In this paper, we propose a data-driven protection framework to address the challenges induced by DERs. Considering the limited available data under fault conditions, we adopt the support vector data description (SVDD) method, a commonly used one-class classifier, for distribution system fault detection. The proposed method is tested under the IEEE 123-node test feeder and simulation results show that our proposed SVDD-based fault detection method significantly improves the robustness and resilience against DERs in comparison with conventional protection systems.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127035411","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646466
Xiao Hu, Xiaoqiang Guo, Zhuqing Jiang, Yun Zhou, Zixuan Yang
Person re-identification (re-id) aims to match person images captured in non-overlapping camera views. Convolutional Neural Network (CNN) has been verified to be powerful in pedestrian feature extraction. However, the CNN features focus more on global visual information, which are sensitive to environmental variations. In comparison, attribute features contain semantic information and prove to be more stable to cross-view appearance changes. In this paper, we present a novel network which leverages high-level semantic attributes to enhance pedestrian descriptors. By introducing hand-crafted multi-colorspaces and texture information to refine CNN features, we acquire a more invariant and reliable feature representation for attribute prediction. The attribute-based stream is further embedded into a part-based CNN branch for re-id. This part-based CNN is trained with a weighted integration of multi-part identification losses. Experiments on two public datasets demonstrate significant performance improvements of our method over state of the arts.
{"title":"PERSON RE-IDENTIFICATION BY REFINED ATTRIBUTE PREDICTION AND WEIGHTED MULTI-PART CONSTRAINTS","authors":"Xiao Hu, Xiaoqiang Guo, Zhuqing Jiang, Yun Zhou, Zixuan Yang","doi":"10.1109/GlobalSIP.2018.8646466","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646466","url":null,"abstract":"Person re-identification (re-id) aims to match person images captured in non-overlapping camera views. Convolutional Neural Network (CNN) has been verified to be powerful in pedestrian feature extraction. However, the CNN features focus more on global visual information, which are sensitive to environmental variations. In comparison, attribute features contain semantic information and prove to be more stable to cross-view appearance changes. In this paper, we present a novel network which leverages high-level semantic attributes to enhance pedestrian descriptors. By introducing hand-crafted multi-colorspaces and texture information to refine CNN features, we acquire a more invariant and reliable feature representation for attribute prediction. The attribute-based stream is further embedded into a part-based CNN branch for re-id. This part-based CNN is trained with a weighted integration of multi-part identification losses. Experiments on two public datasets demonstrate significant performance improvements of our method over state of the arts.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"26 9","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132899806","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646518
S. Pratiher, S. Chattoraj, Rajdeep Mukherjee
A novel non-stationarity visualization tool known as StationPlot is developed for deciphering the chaotic behavior of a dynamical time series. A family of analytic measures enumerating geometrical aspects of the non-stationarity & degree of variability is formulated by convex hull geometry (CHG) on StationPlot. In the Euclidean space, both trend-stationary (TS) & difference-stationary (DS) perturbations are comprehended by the asymmetric structure of StationPlot’s region of interest (ROI). The proposed method is experimentally validated using EEG signals, where it comprehend the relative temporal evolution of neural dynamics & its non-stationary morphology, thereby exemplifying its diagnostic competence for seizure activity (SA) detection. Experimental results & analysis-of-Variance (ANOVA) on the extracted CHG features demonstrates better classification performances as compared to the existing shallow feature based state-of-the-art & validates its efficacy as geometry-rich discriminative descriptors for signal processing applications.
{"title":"StationPlot: A New Non-stationarity Quantification Tool for Detection of Epileptic Seizures","authors":"S. Pratiher, S. Chattoraj, Rajdeep Mukherjee","doi":"10.1109/GlobalSIP.2018.8646518","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646518","url":null,"abstract":"A novel non-stationarity visualization tool known as StationPlot is developed for deciphering the chaotic behavior of a dynamical time series. A family of analytic measures enumerating geometrical aspects of the non-stationarity & degree of variability is formulated by convex hull geometry (CHG) on StationPlot. In the Euclidean space, both trend-stationary (TS) & difference-stationary (DS) perturbations are comprehended by the asymmetric structure of StationPlot’s region of interest (ROI). The proposed method is experimentally validated using EEG signals, where it comprehend the relative temporal evolution of neural dynamics & its non-stationary morphology, thereby exemplifying its diagnostic competence for seizure activity (SA) detection. Experimental results & analysis-of-Variance (ANOVA) on the extracted CHG features demonstrates better classification performances as compared to the existing shallow feature based state-of-the-art & validates its efficacy as geometry-rich discriminative descriptors for signal processing applications.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132041783","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646592
Xiao Hu, Xiaoqiang Guo, Zhuqing Jiang, Yun Zhou, Zixuan Yang
Person Re-Identification (ReID) aims to match people across disjoint camera views. Feature representation and matching are two critical components in person ReID task. In this paper, we introduce a region-partition based bilinear network (RPBi-Net), aiming to capture both global and local information simultaneously. Firstly, a novel Part Box Estimation (PBE) sub-network is embedded to operate region partition on original image. Considering the different importance of human parts, we propose a weighted region partition loss when learning PBE. Secondly, a two stream convolutional neural network is built to learn high-level feature representation from both the whole and partitioned human body. Finally, the learned local and global features are fused in a compact bilinear way, so as to acquire a final descriptor for matching pedestrians. Experimental validation on three benchmark datasets, i.e., CUHK01, CUHK03, Market1501, demonstrates that our model significantly outperforms the state-of-the-art methods.
{"title":"REGION-PARTITION BASED BILINEAR FUSION NETWORK FOR PERSON RE-IDENTIFICATION","authors":"Xiao Hu, Xiaoqiang Guo, Zhuqing Jiang, Yun Zhou, Zixuan Yang","doi":"10.1109/GlobalSIP.2018.8646592","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646592","url":null,"abstract":"Person Re-Identification (ReID) aims to match people across disjoint camera views. Feature representation and matching are two critical components in person ReID task. In this paper, we introduce a region-partition based bilinear network (RPBi-Net), aiming to capture both global and local information simultaneously. Firstly, a novel Part Box Estimation (PBE) sub-network is embedded to operate region partition on original image. Considering the different importance of human parts, we propose a weighted region partition loss when learning PBE. Secondly, a two stream convolutional neural network is built to learn high-level feature representation from both the whole and partitioned human body. Finally, the learned local and global features are fused in a compact bilinear way, so as to acquire a final descriptor for matching pedestrians. Experimental validation on three benchmark datasets, i.e., CUHK01, CUHK03, Market1501, demonstrates that our model significantly outperforms the state-of-the-art methods.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131006125","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646662
H. Salehinejad, S. Naqvi, E. Colak, J. Barfett, S. Valaee
Efficient training of supervised deep learning models for semantic segmentation requires a massive volume of annotated data. In this paper, we propose an unsupervised semantic segmentation method through the application of a radial transform method in the polar coordinate system to unannotated images. This method generates radial transformed images up to the number of pixels in the input image. Each generated image corresponds to a pixel in the original image, which is a spatial representation of the selected pixel with respect to other pixels, in the polar coordinate system. A dimension reduction method, such as a convolutional autoencoder (CAE), can extract features of these representations for clustering and later, labeling the original image by mapping the labels back to the original image. The advantage of the proposed radial transform technique is that it generates a massive number of training images by sampling pixels in the polar coordinate system from a very limited number of original images in the Cartesian coordinate system. The proposed approach achieved 88.20% accuracy in pixel-level segmentation of left kidney, right kidney, and non-kidney pixels in contrast-enhanced computed tomography (CT) images.
{"title":"UNSUPERVISED SEMANTIC SEGMENTATION OF KIDNEYS USING RADIAL TRANSFORM SAMPLING ON LIMITED IMAGES","authors":"H. Salehinejad, S. Naqvi, E. Colak, J. Barfett, S. Valaee","doi":"10.1109/GlobalSIP.2018.8646662","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646662","url":null,"abstract":"Efficient training of supervised deep learning models for semantic segmentation requires a massive volume of annotated data. In this paper, we propose an unsupervised semantic segmentation method through the application of a radial transform method in the polar coordinate system to unannotated images. This method generates radial transformed images up to the number of pixels in the input image. Each generated image corresponds to a pixel in the original image, which is a spatial representation of the selected pixel with respect to other pixels, in the polar coordinate system. A dimension reduction method, such as a convolutional autoencoder (CAE), can extract features of these representations for clustering and later, labeling the original image by mapping the labels back to the original image. The advantage of the proposed radial transform technique is that it generates a massive number of training images by sampling pixels in the polar coordinate system from a very limited number of original images in the Cartesian coordinate system. The proposed approach achieved 88.20% accuracy in pixel-level segmentation of left kidney, right kidney, and non-kidney pixels in contrast-enhanced computed tomography (CT) images.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"116 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121620437","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646455
Yusuke Nakahara, Takuro Yamaguchi, M. Ikehara
In this paper, we propose a single image super-resolution with limited number of filters based on RAISR. RAISR is well known as rapid and accurate super-resolution method which utilizes 864 filters for upscaling. This super-resolution idea utilizes the filter learned with sufficient training set. To get low cost of calculation and comparable image quality with other highly accurate super-resolution methods, the patch of input image is classified into classes by simple hash calculation. Then, the high quality version of this patch is generated by applying the filter to low resolution patches. In our method, only 18 filters can make high resolution images by using simple geometric conversion and rotation conversion while keeping the accuracy and runtime of RAISR.
{"title":"SINGLE IMAGE SUPER-RESOLUTION WITH LIMITED NUMBER OF FILTERS","authors":"Yusuke Nakahara, Takuro Yamaguchi, M. Ikehara","doi":"10.1109/GlobalSIP.2018.8646455","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646455","url":null,"abstract":"In this paper, we propose a single image super-resolution with limited number of filters based on RAISR. RAISR is well known as rapid and accurate super-resolution method which utilizes 864 filters for upscaling. This super-resolution idea utilizes the filter learned with sufficient training set. To get low cost of calculation and comparable image quality with other highly accurate super-resolution methods, the patch of input image is classified into classes by simple hash calculation. Then, the high quality version of this patch is generated by applying the filter to low resolution patches. In our method, only 18 filters can make high resolution images by using simple geometric conversion and rotation conversion while keeping the accuracy and runtime of RAISR.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116703464","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646437
Y. Zhang, Ki Hyun Won, S. Son, A. Siemion, S. Croft
The Search for Extra-terrestrial Intelligence (SETI) aims to find technological signals of extra-solar origin. Radio frequency SETI is characterized by large unlabeled datasets and complex interference environment. The infinite possibilities of potential signal types require generalizable signal processing techniques with little human supervision. We present a generative model of self-supervised deep learning that can be used for anomaly detection and spatial filtering. We develop and evaluate our approach on spectrograms containing narrowband signals collected by Breakthrough Listen at the Green Bank telescope. The proposed approach is not meant to replace current narrowband searches but to demonstrate the potential to generalize to other signal types.
{"title":"SELF-SUPERVISED ANOMALY DETECTION FOR NARROWBAND SETI","authors":"Y. Zhang, Ki Hyun Won, S. Son, A. Siemion, S. Croft","doi":"10.1109/GlobalSIP.2018.8646437","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646437","url":null,"abstract":"The Search for Extra-terrestrial Intelligence (SETI) aims to find technological signals of extra-solar origin. Radio frequency SETI is characterized by large unlabeled datasets and complex interference environment. The infinite possibilities of potential signal types require generalizable signal processing techniques with little human supervision. We present a generative model of self-supervised deep learning that can be used for anomaly detection and spatial filtering. We develop and evaluate our approach on spectrograms containing narrowband signals collected by Breakthrough Listen at the Green Bank telescope. The proposed approach is not meant to replace current narrowband searches but to demonstrate the potential to generalize to other signal types.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121701304","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646598
Martijn Boussé, L. D. Lathauwer
By exploiting the intrinsic structure and/or sparsity of the system coefficients in large-scale system identification, one can enable efficient processing. In this paper, we employ this strategy for large-scale single-input multiple-output autoregressive system identification by assuming the coefficients can be well approximated by Kronecker products of smaller vectors. We show that the identification problem can be refor-mulated as the computation of a Kronecker product equation, allowing one to use optimization-based and algebraic solvers.
{"title":"LARGE-SCALE AUTOREGRESSIVE SYSTEM IDENTIFICATION USING KRONECKER PRODUCT EQUATIONS","authors":"Martijn Boussé, L. D. Lathauwer","doi":"10.1109/GlobalSIP.2018.8646598","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646598","url":null,"abstract":"By exploiting the intrinsic structure and/or sparsity of the system coefficients in large-scale system identification, one can enable efficient processing. In this paper, we employ this strategy for large-scale single-input multiple-output autoregressive system identification by assuming the coefficients can be well approximated by Kronecker products of smaller vectors. We show that the identification problem can be refor-mulated as the computation of a Kronecker product equation, allowing one to use optimization-based and algebraic solvers.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121726767","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646390
C. Wang, Yang Song, Wee Peng Tay
We consider the problem of preserving the privacy of a set of private parameters while allowing inference of a set of public parameters based on observations from sensors in a network. We assume that the public and private parameters are correlated with the sensor observations via a linear model. We define the utility loss and privacy gain functions based on the Cramér-Rao lower bounds for estimating the public and private parameters, respectively. Our goal is to minimize the utility loss while ensuring that the privacy gain is no less than a predefined privacy gain threshold, by allowing each sensor to perturb its own observation before sending it to the fusion center. We propose methods to determine the amount of noise each sensor needs to add to its observation under the cases where prior information is available or unavailable.
{"title":"PRESERVING PARAMETER PRIVACY IN SENSOR NETWORKS","authors":"C. Wang, Yang Song, Wee Peng Tay","doi":"10.1109/GlobalSIP.2018.8646390","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646390","url":null,"abstract":"We consider the problem of preserving the privacy of a set of private parameters while allowing inference of a set of public parameters based on observations from sensors in a network. We assume that the public and private parameters are correlated with the sensor observations via a linear model. We define the utility loss and privacy gain functions based on the Cramér-Rao lower bounds for estimating the public and private parameters, respectively. Our goal is to minimize the utility loss while ensuring that the privacy gain is no less than a predefined privacy gain threshold, by allowing each sensor to perturb its own observation before sending it to the fusion center. We propose methods to determine the amount of noise each sensor needs to add to its observation under the cases where prior information is available or unavailable.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"165 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133524789","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}
Pub Date : 2018-11-01DOI: 10.1109/GlobalSIP.2018.8646327
Aditya Raikar, Sourya Basu, R. Hegde
Single channel speech de-reverberation and de-noising is a challenging problem, since directional information is not available in a single channel when compared to multi-channel approaches. Several deep neural network (DNN) based solutions have been proposed in the recent past to solve this problem. These solutions are sequential and de-reverberate the signal after denoising. Additionally these solutions have not utilized the maximum a posteriori (MAP) method which requires the knowledge of the prior. In this work a MAP method is proposed to solve the speech de-reverberation and de-noising problem jointly. A half quadratic splitting (HQS) method is used to solve the joint MAP problem in a DNN framework by splitting it into two minimization problems. The deep prior is modeled using a latent variable and obtained using an iterative method. The performance of the proposed method is illustrated using subjective and objective measures. Experiments on continuous speech recognition are also used to demonstrate the significance of this method.
{"title":"SINGLE CHANNEL JOINT SPEECH DEREVERBERATION AND DENOISING USING DEEP PRIORS","authors":"Aditya Raikar, Sourya Basu, R. Hegde","doi":"10.1109/GlobalSIP.2018.8646327","DOIUrl":"https://doi.org/10.1109/GlobalSIP.2018.8646327","url":null,"abstract":"Single channel speech de-reverberation and de-noising is a challenging problem, since directional information is not available in a single channel when compared to multi-channel approaches. Several deep neural network (DNN) based solutions have been proposed in the recent past to solve this problem. These solutions are sequential and de-reverberate the signal after denoising. Additionally these solutions have not utilized the maximum a posteriori (MAP) method which requires the knowledge of the prior. In this work a MAP method is proposed to solve the speech de-reverberation and de-noising problem jointly. A half quadratic splitting (HQS) method is used to solve the joint MAP problem in a DNN framework by splitting it into two minimization problems. The deep prior is modeled using a latent variable and obtained using an iterative method. The performance of the proposed method is illustrated using subjective and objective measures. Experiments on continuous speech recognition are also used to demonstrate the significance of this method.","PeriodicalId":119131,"journal":{"name":"2018 IEEE Global Conference on Signal and Information Processing (GlobalSIP)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134147658","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：