Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287617
Zois Boukouvalas, Monica Puerto, D. Elton, Peter W. Chung, M. Fuge
Due to its high computational speed and accuracy compared to ab-initio quantum chemistry and forcefield modeling, the prediction of molecular properties using machine learning has received great attention in the fields of materials design and drug discovery. A main ingredient required for machine learning is a training dataset consisting of molecular features—for example fingerprint bits, chemical descriptors, etc. that adequately characterize the corresponding molecules. However, choosing features for any application is highly non-trivial, since no "universal" method for feature selection exists. In this work, we propose a data fusion framework that uses Independent Vector Analysis to uncover underlying complementary information contained in different molecular featurization methods. Our approach takes an arbitrary number of individual feature vectors and generates a low dimensional set of features—molecular signatures—that can be used for the prediction of molecular properties and for knowledge discovery. We demonstrate this on a small and diverse dataset consisting of energetic compounds for the prediction of several energetic properties as well as for demonstrating how to provide insights onto the relationships between molecular structures and properties.
{"title":"Independent Vector Analysis for Molecular Data Fusion: Application to Property Prediction and Knowledge Discovery of Energetic Materials","authors":"Zois Boukouvalas, Monica Puerto, D. Elton, Peter W. Chung, M. Fuge","doi":"10.23919/Eusipco47968.2020.9287617","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287617","url":null,"abstract":"Due to its high computational speed and accuracy compared to ab-initio quantum chemistry and forcefield modeling, the prediction of molecular properties using machine learning has received great attention in the fields of materials design and drug discovery. A main ingredient required for machine learning is a training dataset consisting of molecular features—for example fingerprint bits, chemical descriptors, etc. that adequately characterize the corresponding molecules. However, choosing features for any application is highly non-trivial, since no \"universal\" method for feature selection exists. In this work, we propose a data fusion framework that uses Independent Vector Analysis to uncover underlying complementary information contained in different molecular featurization methods. Our approach takes an arbitrary number of individual feature vectors and generates a low dimensional set of features—molecular signatures—that can be used for the prediction of molecular properties and for knowledge discovery. We demonstrate this on a small and diverse dataset consisting of energetic compounds for the prediction of several energetic properties as well as for demonstrating how to provide insights onto the relationships between molecular structures and properties.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"20 1","pages":"1030-1034"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81101048","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287568
G. Tzagkarakis, F. Maurer, T. Dionysopoulos
Quantifying risk is pivotal for every financial institution, with the temporal dimension being the key aspect for all the well-established risk measures. However, exploiting the frequency information conveyed by financial data, could yield improved insights about the inherent risk evolution in a joint time-frequency fashion. Nevertheless, the great majority of risk managers make no explicit distinction between the information captured by patterns of different frequency content, while relying on the full time-resolution data, regardless of the trading horizon. To address this problem, a novel value-at-risk (VaR) quantification method is proposed, which combines nonlinearly the time-evolving energy profile of returns series at multiple frequency scales, determined by the predefined trading horizon. Most importantly, our proposed method can be coupled with any quantile-based risk measure to enhance its performance. Experimental evaluation with real data reveals an increased robustness of our method in efficiently controlling under-/over-estimated VaR values.
{"title":"Nonparametric Adaptive Value-at-Risk Quantification Based on the Multiscale Energy Distribution of Asset Returns","authors":"G. Tzagkarakis, F. Maurer, T. Dionysopoulos","doi":"10.23919/Eusipco47968.2020.9287568","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287568","url":null,"abstract":"Quantifying risk is pivotal for every financial institution, with the temporal dimension being the key aspect for all the well-established risk measures. However, exploiting the frequency information conveyed by financial data, could yield improved insights about the inherent risk evolution in a joint time-frequency fashion. Nevertheless, the great majority of risk managers make no explicit distinction between the information captured by patterns of different frequency content, while relying on the full time-resolution data, regardless of the trading horizon. To address this problem, a novel value-at-risk (VaR) quantification method is proposed, which combines nonlinearly the time-evolving energy profile of returns series at multiple frequency scales, determined by the predefined trading horizon. Most importantly, our proposed method can be coupled with any quantile-based risk measure to enhance its performance. Experimental evaluation with real data reveals an increased robustness of our method in efficiently controlling under-/over-estimated VaR values.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"15 1","pages":"2393-2397"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82517821","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287492
Huanyu Zuo, P. Samarasinghe, T. Abhayapala
Intensity based soundfield reproduction methods are shown to provide impressive human perception of sound localization. However, most of the previous works in this domain either focus on a single sweet spot for the listener, or are constrained to a regular loudspeaker geometry, which is difficult to implement in real-world applications. This paper addresses both of the above challenges. We propose an intensity matching technique to optimally reproduce sound intensity at multiple sweet spots using an irregular loudspeaker array. The performance of the proposed method is evaluated by comparing it with the pressure and velocity matching method through numerical simulations and perceptual experiments. The results show that the proposed method has an improved performance.
{"title":"Intensity Based Soundfield Reproduction over Multiple Sweet Spots Using an Irregular Loudspeaker Array","authors":"Huanyu Zuo, P. Samarasinghe, T. Abhayapala","doi":"10.23919/Eusipco47968.2020.9287492","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287492","url":null,"abstract":"Intensity based soundfield reproduction methods are shown to provide impressive human perception of sound localization. However, most of the previous works in this domain either focus on a single sweet spot for the listener, or are constrained to a regular loudspeaker geometry, which is difficult to implement in real-world applications. This paper addresses both of the above challenges. We propose an intensity matching technique to optimally reproduce sound intensity at multiple sweet spots using an irregular loudspeaker array. The performance of the proposed method is evaluated by comparing it with the pressure and velocity matching method through numerical simulations and perceptual experiments. The results show that the proposed method has an improved performance.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"93 1","pages":"486-490"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80025817","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287771
Robbe Van Rompaey, M. Moonen
This paper presents a distributed adaptive algorithm for node-specific sound zoning in a wireless acoustic sensor and actuator network (WASAN), based on a network-wide acoustic contrast control (ACC) method. The goal of the ACC method is to simultaneously create node-specific zones with high signal power (bright zones) while minimizing power leakage in other node-specific zones (dark zones). To obtain this, a network-wide objective involving the acoustic coupling between all the loudspeakers and microphones in the WASAN is proposed where the optimal solution is based on a centralized generalized eigenvalue decomposition (GEVD). To allow for distributed processing, a gradient based GEVD algorithm is first proposed that minimizes the same objective. This algorithm can then be modified to allow for a fully distributed implementation, involving in-network summations and simple local processing. The algorithm is referred to as the distributed adaptive gradient based ACC algorithm (DAGACC). The proposed algorithm outperforms the non-cooperative distributed solution after only a few iterations and converges to the centralized solution, as illustrated by computer simulations.
{"title":"Distributed Adaptive Acoustic Contrast Control for Node-specific Sound Zoning in a Wireless Acoustic Sensor and Actuator Network","authors":"Robbe Van Rompaey, M. Moonen","doi":"10.23919/Eusipco47968.2020.9287771","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287771","url":null,"abstract":"This paper presents a distributed adaptive algorithm for node-specific sound zoning in a wireless acoustic sensor and actuator network (WASAN), based on a network-wide acoustic contrast control (ACC) method. The goal of the ACC method is to simultaneously create node-specific zones with high signal power (bright zones) while minimizing power leakage in other node-specific zones (dark zones). To obtain this, a network-wide objective involving the acoustic coupling between all the loudspeakers and microphones in the WASAN is proposed where the optimal solution is based on a centralized generalized eigenvalue decomposition (GEVD). To allow for distributed processing, a gradient based GEVD algorithm is first proposed that minimizes the same objective. This algorithm can then be modified to allow for a fully distributed implementation, involving in-network summations and simple local processing. The algorithm is referred to as the distributed adaptive gradient based ACC algorithm (DAGACC). The proposed algorithm outperforms the non-cooperative distributed solution after only a few iterations and converges to the centralized solution, as illustrated by computer simulations.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"30 4 1","pages":"481-485"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81588316","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287805
T. Huuhtanen, A. Lankinen, Alexander Jung
An emerging class of applications uses sensing surfaces, where sensor data is collected from a 2-dimensional surface covering a large spatial area. Sensing surface applications range from observing human activity to detecting failures of construction materials. Electrical impedance tomography (EIT) is an imaging technology, which has been successfully applied to imaging in several important application domains such as medicine, geophysics, and process industry. EIT is a low-cost technology offering high temporal resolution, which makes it a potential technology sensing surfaces. In this paper, we evaluate the applicability of EIT algorithms for tracking a small moving object on a 2D sensing surface. We compare standard EIT algorithms for this purpose and develop a method which models the movement of a small target on a sensing surface using hidden Markov models (HMM). Existing EIT methods are geared towards high image quality instead of smooth target trajectories, which makes them suboptimal for target tracking. Numerical experiments indicate that our proposed method outperforms existing EIT methods in target tracking accuracy.
{"title":"Target Tracking on Sensing Surface with Electrical Impedance Tomography","authors":"T. Huuhtanen, A. Lankinen, Alexander Jung","doi":"10.23919/Eusipco47968.2020.9287805","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287805","url":null,"abstract":"An emerging class of applications uses sensing surfaces, where sensor data is collected from a 2-dimensional surface covering a large spatial area. Sensing surface applications range from observing human activity to detecting failures of construction materials. Electrical impedance tomography (EIT) is an imaging technology, which has been successfully applied to imaging in several important application domains such as medicine, geophysics, and process industry. EIT is a low-cost technology offering high temporal resolution, which makes it a potential technology sensing surfaces. In this paper, we evaluate the applicability of EIT algorithms for tracking a small moving object on a 2D sensing surface. We compare standard EIT algorithms for this purpose and develop a method which models the movement of a small target on a sensing surface using hidden Markov models (HMM). Existing EIT methods are geared towards high image quality instead of smooth target trajectories, which makes them suboptimal for target tracking. Numerical experiments indicate that our proposed method outperforms existing EIT methods in target tracking accuracy.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"30 1","pages":"1817-1821"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87344806","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287592
Payam Shahsavari Baboukani, C. Graversen, Jan Østergaard
It is well-known that estimation of the directed dependency between high-dimensional data sequences suffers from the "curse of dimensionality" problem. To reduce the dimensionality of the data, and thereby improve the accuracy of the estimation, we propose a new progressive input variable selection technique. Specifically, in each iteration, the remaining input variables are ranked according to a weighted sum of the amount of new information provided by the variable and the variable’s prediction accuracy. Then, the highest ranked variable is included, if it is significant enough to improve the accuracy of the prediction. A simulation study on synthetic non-linear autoregressive and Henon maps data, shows a significant improvement over existing estimator, especially in the case of small amounts of high-dimensional and highly correlated data.
{"title":"Estimation of Directed Dependencies in Time Series Using Conditional Mutual Information and Non-linear Prediction","authors":"Payam Shahsavari Baboukani, C. Graversen, Jan Østergaard","doi":"10.23919/Eusipco47968.2020.9287592","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287592","url":null,"abstract":"It is well-known that estimation of the directed dependency between high-dimensional data sequences suffers from the \"curse of dimensionality\" problem. To reduce the dimensionality of the data, and thereby improve the accuracy of the estimation, we propose a new progressive input variable selection technique. Specifically, in each iteration, the remaining input variables are ranked according to a weighted sum of the amount of new information provided by the variable and the variable’s prediction accuracy. Then, the highest ranked variable is included, if it is significant enough to improve the accuracy of the prediction. A simulation study on synthetic non-linear autoregressive and Henon maps data, shows a significant improvement over existing estimator, especially in the case of small amounts of high-dimensional and highly correlated data.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"5 1","pages":"2388-2392"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88327027","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287873
Harald Bârzan, V. V. Moca, Ana-Maria Ichim, R. Muresan
The Continuous Wavelet Transform (CWT) provides a multi-resolution representation of a signal by scaling a mother wavelet and convolving it with the signal. The scalogram (squared modulus of the CWT) then represents the spread of the signal's energy as a function of time and scale. The scalogram has constant relative temporal resolution but, as the scale is compressed (frequency increased), it loses frequency resolution. To compensate for this, the recently-introduced superlets geometrically combine a set of wavelets with increasing frequency resolution to achieve time-frequency super-resolution. The number of wavelets in the set is called the order of the superlet and was initially defined as an integer number. This creates a series of issues when adaptive superlets are implemented, i.e. superlets whose order depends on frequency. In particular, adaptive superlets generate representations that suffer from "banding" because the order is adjusted in discrete steps as the frequency increases. Here, by relying on the weighted geometric mean, we introduce fractional superlets, which allow the order to be a fractional number. We show that fractional adaptive superlets provide high-resolution representations that are smooth across the entire spectrum and are clearly superior to representations based on the discrete adaptive superlets.
{"title":"Fractional Superlets","authors":"Harald Bârzan, V. V. Moca, Ana-Maria Ichim, R. Muresan","doi":"10.23919/Eusipco47968.2020.9287873","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287873","url":null,"abstract":"The Continuous Wavelet Transform (CWT) provides a multi-resolution representation of a signal by scaling a mother wavelet and convolving it with the signal. The scalogram (squared modulus of the CWT) then represents the spread of the signal's energy as a function of time and scale. The scalogram has constant relative temporal resolution but, as the scale is compressed (frequency increased), it loses frequency resolution. To compensate for this, the recently-introduced superlets geometrically combine a set of wavelets with increasing frequency resolution to achieve time-frequency super-resolution. The number of wavelets in the set is called the order of the superlet and was initially defined as an integer number. This creates a series of issues when adaptive superlets are implemented, i.e. superlets whose order depends on frequency. In particular, adaptive superlets generate representations that suffer from \"banding\" because the order is adjusted in discrete steps as the frequency increases. Here, by relying on the weighted geometric mean, we introduce fractional superlets, which allow the order to be a fractional number. We show that fractional adaptive superlets provide high-resolution representations that are smooth across the entire spectrum and are clearly superior to representations based on the discrete adaptive superlets.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"31 1","pages":"2220-2224"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88303668","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287538
A. Lawal, Qadri Mayyala, A. Zerguine, Azeddine Beghdadi
This work presents a method for salt dome detection in seismic images based on a Context-Aware Saliency (CAS) detection model. Seismic data can easily add up to hundred of gigabytes and terabytes in size. However, the key features or structural information that are of interest to the seismic interpreters are quite few. These features include salt domes, fault and other geological features that have the potential of indicating the presence of oil reservoir. A new method for extracting the most perceptual relevant features in seismic images based on the CAS model is proposed. The efficiency of this method in detecting the most salient structures in a seismic image such as salt dome is demonstrated through a series of experiment on real data set with various spatial contents.
{"title":"Salt Dome Detection Using Context-Aware Saliency","authors":"A. Lawal, Qadri Mayyala, A. Zerguine, Azeddine Beghdadi","doi":"10.23919/Eusipco47968.2020.9287538","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287538","url":null,"abstract":"This work presents a method for salt dome detection in seismic images based on a Context-Aware Saliency (CAS) detection model. Seismic data can easily add up to hundred of gigabytes and terabytes in size. However, the key features or structural information that are of interest to the seismic interpreters are quite few. These features include salt domes, fault and other geological features that have the potential of indicating the presence of oil reservoir. A new method for extracting the most perceptual relevant features in seismic images based on the CAS model is proposed. The efficiency of this method in detecting the most salient structures in a seismic image such as salt dome is demonstrated through a series of experiment on real data set with various spatial contents.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"21 1","pages":"1906-1910"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89873611","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287832
Sisi Shi, Andrew Busch, K. Paliwal, T. Fickenscher
This paper investigates the use of short-time Discrete Cosine Transform (DCT) for speech enhancement. We denote the absolute values and signs of the DCT spectral coefficients as the Absolute Spectrum (AS) and Polarity Spectrum (PoS), respectively. We theoretically show that the noisy PoS is the best estimate of the original, under the constrained MMSE criterion. To verify this experimentally, the effect of using the noisy PoS for signal resynthesis is analysed through objective and subjective measures. The results show that when the Instantaneous SNR (ISNR) is above 0 dB, deemed as perfect, recovery of the original speech signal can be obtained only by modifying the DCT absolute spectrum. However, an accurate DFT Phase Spectrum (PhS) estimation might be required to achieve the same improvement in perceived speech quality. When the perceived quality is measured against the Segmental SNR (SSNR), it shows the PoS is more capable to conserve the speech quality than the PhS for the same level of global distortion. The results show that the noisy PoS can be used as an estimate of the clean PoS without perceivable degradation in speech quality, only if the ISNR of the noisy speech signal is above 0 dB or the SSNR is above 10.5 dB.
{"title":"On The Use of Discrete Cosine Transform Polarity Spectrum in Speech Enhancement","authors":"Sisi Shi, Andrew Busch, K. Paliwal, T. Fickenscher","doi":"10.23919/Eusipco47968.2020.9287832","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287832","url":null,"abstract":"This paper investigates the use of short-time Discrete Cosine Transform (DCT) for speech enhancement. We denote the absolute values and signs of the DCT spectral coefficients as the Absolute Spectrum (AS) and Polarity Spectrum (PoS), respectively. We theoretically show that the noisy PoS is the best estimate of the original, under the constrained MMSE criterion. To verify this experimentally, the effect of using the noisy PoS for signal resynthesis is analysed through objective and subjective measures. The results show that when the Instantaneous SNR (ISNR) is above 0 dB, deemed as perfect, recovery of the original speech signal can be obtained only by modifying the DCT absolute spectrum. However, an accurate DFT Phase Spectrum (PhS) estimation might be required to achieve the same improvement in perceived speech quality. When the perceived quality is measured against the Segmental SNR (SSNR), it shows the PoS is more capable to conserve the speech quality than the PhS for the same level of global distortion. The results show that the noisy PoS can be used as an estimate of the clean PoS without perceivable degradation in speech quality, only if the ISNR of the noisy speech signal is above 0 dB or the SSNR is above 10.5 dB.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"122 1","pages":"421-425"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89873421","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 : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287512
Moe Takada, Shogo Seki, Patrick Lumban Tobing, T. Toda
We propose a semi-supervised method for enhancing and suppressing self-produced speech recorded with wearable air- and body-conductive microphones. Body-conducted signals are robust against external noise and predominantly contain self-produced speech. As a result, these signals provide informative acoustical clues when estimating a linear filter to separate a mixed signal into self-produced speech and background noise. In a previous study, we proposed a blind source separation method for handling air- and body-conducted signals as a multi-channel signal. While our previously proposed method demonstrated the superior performance that can be achieved by using air- and body-conducted signals in comparison to using only air-conducted signals, the enhanced and suppressed air-conducted signals tended to be contaminated with the acoustical characteristics of the body-conducted signals due to the nonlinear relationship between these signals. To address this issue, in this paper, we introduce a new source model which takes into consideration the correspondence between these signals and incorporates them within a semi-supervised framework. Our experimental results reveal that this new method alleviates the negative effects of using the acoustical characteristics of the body-conducted signals, outperforming our previously proposed method, as well as conventional methods, under a semi-supervised condition.
{"title":"Semi-Supervised Enhancement and Suppression of Self-Produced Speech Using Correspondence between Air- and Body-Conducted Signals","authors":"Moe Takada, Shogo Seki, Patrick Lumban Tobing, T. Toda","doi":"10.23919/Eusipco47968.2020.9287512","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287512","url":null,"abstract":"We propose a semi-supervised method for enhancing and suppressing self-produced speech recorded with wearable air- and body-conductive microphones. Body-conducted signals are robust against external noise and predominantly contain self-produced speech. As a result, these signals provide informative acoustical clues when estimating a linear filter to separate a mixed signal into self-produced speech and background noise. In a previous study, we proposed a blind source separation method for handling air- and body-conducted signals as a multi-channel signal. While our previously proposed method demonstrated the superior performance that can be achieved by using air- and body-conducted signals in comparison to using only air-conducted signals, the enhanced and suppressed air-conducted signals tended to be contaminated with the acoustical characteristics of the body-conducted signals due to the nonlinear relationship between these signals. To address this issue, in this paper, we introduce a new source model which takes into consideration the correspondence between these signals and incorporates them within a semi-supervised framework. Our experimental results reveal that this new method alleviates the negative effects of using the acoustical characteristics of the body-conducted signals, outperforming our previously proposed method, as well as conventional methods, under a semi-supervised condition.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"1 1","pages":"456-460"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85250606","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：