Pub Date : 2021-01-24DOI: 10.23919/Eusipco47968.2020.9287494
Pedro Marinho R. de Oliveira, V. Zarzoso, C. A. R. Fernandes
Atrial fibrillation (AF) is the most frequent cardiac arrhythmia diagnosed in clinical practice, identified by an uncoordinated and irregular atrial depolarization. However, its electrophysiological mechanisms are still not clearly understood, increasing the intensive clinical research into this challenging cardiac condition in the past few years. The noninvasive extraction of the atrial activity (AA) from multi-lead electrocardiogram (ECG) recordings by signal processing techniques has helped in better understanding this complex arrhythmia. In particular, tensor decomposition techniques have proven to be powerful tools in this task, overcoming the limitations of matrix factorization methods. Exploring the spatial as well as the temporal diversity of ECG recordings, this contribution puts forward a novel noninvasive AA extraction method that models consecutive AF ECG segments as a coupled block-term tensor decomposition, assuming that they share the same spatial signatures. Experiments on synthetic and real data, the latter acquired from persistent AF patients, validate the proposed coupled tensor approach, which provides satisfactory performance with reduced computational cost.
{"title":"Coupled Tensor Model of Atrial Fibrillation ECG","authors":"Pedro Marinho R. de Oliveira, V. Zarzoso, C. A. R. Fernandes","doi":"10.23919/Eusipco47968.2020.9287494","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287494","url":null,"abstract":"Atrial fibrillation (AF) is the most frequent cardiac arrhythmia diagnosed in clinical practice, identified by an uncoordinated and irregular atrial depolarization. However, its electrophysiological mechanisms are still not clearly understood, increasing the intensive clinical research into this challenging cardiac condition in the past few years. The noninvasive extraction of the atrial activity (AA) from multi-lead electrocardiogram (ECG) recordings by signal processing techniques has helped in better understanding this complex arrhythmia. In particular, tensor decomposition techniques have proven to be powerful tools in this task, overcoming the limitations of matrix factorization methods. Exploring the spatial as well as the temporal diversity of ECG recordings, this contribution puts forward a novel noninvasive AA extraction method that models consecutive AF ECG segments as a coupled block-term tensor decomposition, assuming that they share the same spatial signatures. Experiments on synthetic and real data, the latter acquired from persistent AF patients, validate the proposed coupled tensor approach, which provides satisfactory performance with reduced computational cost.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"42 1","pages":"915-919"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83111205","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/Eusipco47968.2020.9287789
Matheus Lindino, Thiago Bubolz, B. Zatt, D. Palomino, G. Corrêa
Video transcoding for bit rate adaptation has become mandatory for over-the-top applications that deliver multimedia content in heterogeneous environments under different network conditions and user capabilities. As transcoding requires sequentially decoding and re-encoding the video bitstream, the computational cost involved in the process is too high, especially when considering current state-of-the-art codecs, such as the High Efficiency Video Coding (HEVC). This work presents a fast HEVC transcoder for bit rate adaptation based on Prediction Unit (PU) mode inheritance, which uses information gathered from the HEVC decoding process to accelerate PU mode decision in the re-encoding process. Experimental results show that the proposed method achieves an average transrating time reduction of 42% at the cost of a bitrate increase of 0.54%.
{"title":"Low-Complexity HEVC Transrating Based on Prediction Unit Mode Inheritance","authors":"Matheus Lindino, Thiago Bubolz, B. Zatt, D. Palomino, G. Corrêa","doi":"10.23919/Eusipco47968.2020.9287789","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287789","url":null,"abstract":"Video transcoding for bit rate adaptation has become mandatory for over-the-top applications that deliver multimedia content in heterogeneous environments under different network conditions and user capabilities. As transcoding requires sequentially decoding and re-encoding the video bitstream, the computational cost involved in the process is too high, especially when considering current state-of-the-art codecs, such as the High Efficiency Video Coding (HEVC). This work presents a fast HEVC transcoder for bit rate adaptation based on Prediction Unit (PU) mode inheritance, which uses information gathered from the HEVC decoding process to accelerate PU mode decision in the re-encoding process. Experimental results show that the proposed method achieves an average transrating time reduction of 42% at the cost of a bitrate increase of 0.54%.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"14 1","pages":"550-554"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80928905","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/Eusipco47968.2020.9287344
D. Krause, A. Politis, K. Kowalczyk
This paper presents an overview of several approaches to convolutional feature extraction in the context of deep neural network (DNN) based sound source localization. Different ways of processing multichannel audio data in the time-frequency domain using convolutional neural networks (CNNs) are described and tested with the aim to provide a comparative study of their performance. In most considered approaches, models are trained with phase and magnitude components of the Short-Time Fourier Transform (STFT). In addition to state-of-the-art 2D convolutional layers, we investigate several solutions for the processing of 3D matrices containing multichannel complex representation of the microphone signals. The first two proposed approaches are the 3D convolutions and depthwise separable convolutions in which two types of filters are used to exploit information within and between the channels. Note that this paper presents the first application of depthwise separable convolutions in a task of sound source localization. The third approach is based on complex-valued neural networks which allows for performing convolutions directly on complex signal representations. Experiments are conducted using two synthetic datasets containing noise and speech signals recorded using a tetrahedral microphone array. The paper presents the results obtained using all investigated model types and discusses the resulting accuracy and computational complexity in DNN-based source localization.
{"title":"Comparison of Convolution Types in CNN-based Feature Extraction for Sound Source Localization","authors":"D. Krause, A. Politis, K. Kowalczyk","doi":"10.23919/Eusipco47968.2020.9287344","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287344","url":null,"abstract":"This paper presents an overview of several approaches to convolutional feature extraction in the context of deep neural network (DNN) based sound source localization. Different ways of processing multichannel audio data in the time-frequency domain using convolutional neural networks (CNNs) are described and tested with the aim to provide a comparative study of their performance. In most considered approaches, models are trained with phase and magnitude components of the Short-Time Fourier Transform (STFT). In addition to state-of-the-art 2D convolutional layers, we investigate several solutions for the processing of 3D matrices containing multichannel complex representation of the microphone signals. The first two proposed approaches are the 3D convolutions and depthwise separable convolutions in which two types of filters are used to exploit information within and between the channels. Note that this paper presents the first application of depthwise separable convolutions in a task of sound source localization. The third approach is based on complex-valued neural networks which allows for performing convolutions directly on complex signal representations. Experiments are conducted using two synthetic datasets containing noise and speech signals recorded using a tetrahedral microphone array. The paper presents the results obtained using all investigated model types and discusses the resulting accuracy and computational complexity in DNN-based source localization.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"138 1","pages":"820-824"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80988283","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/Eusipco47968.2020.9287482
Amita Giri, L. Kumar, T. Gandhi
ElectroEncephaloGram (EEG) signals based Brain Source Localization (BSL) has been an active area of research. The performance of BSL algorithms is severely degraded in the presence of background interferences. Pre-Whitening (PW) based approach to deal with such interference assumes temporal stationarity of the data which does not hold good for EEG based processing. Null Projection (NP) based approach relaxes the temporal stationarity. However, the strict spatial stationarity of the number of interfering sources is maintained between control state and activity state measurement. In practical scenarios where an interference source that exists only in the control state, and does not appear in activity state, NP based approach removes a higher dimension space from the activity data leading to its poor performance. The proposed Subspace Principal Vector Projection (SPVP) based approach utilizes subspace correlation based common interference statistics and thus relaxing the strict spatial stationarity condition. In particular, SPVP based MUltiple SIgnal Classification (MUSIC) and Linearly Constrained Minimum Variance (LCMV) algorithms are presented for BSL. Simulation and experiment with real EEG data from Physionet dataset involving motor imagery task illustrate the effectiveness of the proposed algorithms in robust BSL with interference suppression.
{"title":"Robust EEG Source Localization Using Subspace Principal Vector Projection Technique","authors":"Amita Giri, L. Kumar, T. Gandhi","doi":"10.23919/Eusipco47968.2020.9287482","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287482","url":null,"abstract":"ElectroEncephaloGram (EEG) signals based Brain Source Localization (BSL) has been an active area of research. The performance of BSL algorithms is severely degraded in the presence of background interferences. Pre-Whitening (PW) based approach to deal with such interference assumes temporal stationarity of the data which does not hold good for EEG based processing. Null Projection (NP) based approach relaxes the temporal stationarity. However, the strict spatial stationarity of the number of interfering sources is maintained between control state and activity state measurement. In practical scenarios where an interference source that exists only in the control state, and does not appear in activity state, NP based approach removes a higher dimension space from the activity data leading to its poor performance. The proposed Subspace Principal Vector Projection (SPVP) based approach utilizes subspace correlation based common interference statistics and thus relaxing the strict spatial stationarity condition. In particular, SPVP based MUltiple SIgnal Classification (MUSIC) and Linearly Constrained Minimum Variance (LCMV) algorithms are presented for BSL. Simulation and experiment with real EEG data from Physionet dataset involving motor imagery task illustrate the effectiveness of the proposed algorithms in robust BSL with interference suppression.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"50 1","pages":"1075-1079"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81038212","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/Eusipco47968.2020.9287366
Ilyes Batatia
This paper proposes an integrated method for recognizing special crystals, called metal-organic frameworks (MOF), in scanning electron microscopy images (SEM). The proposed approach combines two deep learning networks and a dense conditional random field (CRF) to perform image segmentation. A modified Unet-like convolutional neural network (CNN), incorporating dilatation techniques using atrous convolution, is designed to segment cluttered objects in the SEM image. The dense CRF is tailored to enhance object boundaries and recover small objects. The unary energy of the CRF is obtained from the CNN. And the pairwise energy is estimated using mean field approximation. The resulting segmented regions are fed to a fully connected CNN that performs instance recognition. The method has been trained on a dataset of 500 images with 3200 objects from 3 classes. Testing achieves an overall accuracy of 95.7% MOF recognition. The proposed method opens up the possibility for developing automated chemical process monitoring that allows researchers to optimize conditions of MOF synthesis.
{"title":"A Deep Learning Method with CRF for Instance Segmentation of Metal-Organic Frameworks in Scanning Electron Microscopy Images","authors":"Ilyes Batatia","doi":"10.23919/Eusipco47968.2020.9287366","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287366","url":null,"abstract":"This paper proposes an integrated method for recognizing special crystals, called metal-organic frameworks (MOF), in scanning electron microscopy images (SEM). The proposed approach combines two deep learning networks and a dense conditional random field (CRF) to perform image segmentation. A modified Unet-like convolutional neural network (CNN), incorporating dilatation techniques using atrous convolution, is designed to segment cluttered objects in the SEM image. The dense CRF is tailored to enhance object boundaries and recover small objects. The unary energy of the CRF is obtained from the CNN. And the pairwise energy is estimated using mean field approximation. The resulting segmented regions are fed to a fully connected CNN that performs instance recognition. The method has been trained on a dataset of 500 images with 3200 objects from 3 classes. Testing achieves an overall accuracy of 95.7% MOF recognition. The proposed method opens up the possibility for developing automated chemical process monitoring that allows researchers to optimize conditions of MOF synthesis.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"85 1","pages":"625-629"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78831563","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/Eusipco47968.2020.9287380
Federico Borra, Mirco Pezzoli, Luca Comanducci, A. Bernardini, F. Antonacci, S. Tubaro, A. Sarti
The importance of soundfield imaging techniques is expected to further increase in the next few years thanks to the ever-increasing availability of low-cost sensors such as MEMS microphones. When it comes to processing a relevant number of sensor signals, however, the computational load of space-time processing algorithms easily grows to unmanageable levels. The Ray Space Transform (RST) was recently introduced as a promising tool for soundfield analysis. Given the collection of signals captured by a uniform linear array of microphones, the RST allows us to collect and map the directional components of the acoustic field onto a domain called "ray space", where relevant acoustic objects are represented as linear patterns for advanced acoustic analysis and synthesis applications. So far the computational complexity of the RST linearly increases with the number of microphones. In order to alleviate this problem, in this paper we propose an alternative efficient implementation of the RST based on the Non Uniform Fast Fourier Transform.
{"title":"A Fast Ray Space Transform for Wave Field Processing using Acoustic Arrays","authors":"Federico Borra, Mirco Pezzoli, Luca Comanducci, A. Bernardini, F. Antonacci, S. Tubaro, A. Sarti","doi":"10.23919/Eusipco47968.2020.9287380","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287380","url":null,"abstract":"The importance of soundfield imaging techniques is expected to further increase in the next few years thanks to the ever-increasing availability of low-cost sensors such as MEMS microphones. When it comes to processing a relevant number of sensor signals, however, the computational load of space-time processing algorithms easily grows to unmanageable levels. The Ray Space Transform (RST) was recently introduced as a promising tool for soundfield analysis. Given the collection of signals captured by a uniform linear array of microphones, the RST allows us to collect and map the directional components of the acoustic field onto a domain called \"ray space\", where relevant acoustic objects are represented as linear patterns for advanced acoustic analysis and synthesis applications. So far the computational complexity of the RST linearly increases with the number of microphones. In order to alleviate this problem, in this paper we propose an alternative efficient implementation of the RST based on the Non Uniform Fast Fourier Transform.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"26 1","pages":"186-190"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82788335","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/Eusipco47968.2020.9287558
E. Vasileva, Nenad Avramovski, Z. Ivanovski
This paper presents a CNN-based algorithm for detecting package edges in a scene represented with a distance map (range image), trained on a custom dataset of packaging scenarios. The proposed algorithm represents the basis for package recognition for automatic trailer loading/unloading. The main focus of this paper is designing a semantic segmentation CNN model capable of detecting different types of package edges in a distance map containing distance errors characteristic of Time-of-Flight (ToF) scanning, and differentiating box edges from edges belonging to other types of packaging objects (bags, irregular objects, etc.). The proposed CNN is optimized for training with a limited number of samples containing heavily imbalanced classes. Generating a binary mask of edges with 1-pixel thickness from the probability maps outputted from the CNN is achieved through a custom non-maximum suppression-based edge thinning algorithm. The proposed algorithm shows promising results in detecting box edges.
{"title":"Detection of Package Edges in Distance Maps","authors":"E. Vasileva, Nenad Avramovski, Z. Ivanovski","doi":"10.23919/Eusipco47968.2020.9287558","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287558","url":null,"abstract":"This paper presents a CNN-based algorithm for detecting package edges in a scene represented with a distance map (range image), trained on a custom dataset of packaging scenarios. The proposed algorithm represents the basis for package recognition for automatic trailer loading/unloading. The main focus of this paper is designing a semantic segmentation CNN model capable of detecting different types of package edges in a distance map containing distance errors characteristic of Time-of-Flight (ToF) scanning, and differentiating box edges from edges belonging to other types of packaging objects (bags, irregular objects, etc.). The proposed CNN is optimized for training with a limited number of samples containing heavily imbalanced classes. Generating a binary mask of edges with 1-pixel thickness from the probability maps outputted from the CNN is achieved through a custom non-maximum suppression-based edge thinning algorithm. The proposed algorithm shows promising results in detecting box edges.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"24 1","pages":"600-604"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89963709","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/Eusipco47968.2020.9287571
Annika Liebgott, S. Gatidis, V. Vu, Tobias Haueise, K. Nikolaou, Bin Yang
The treatment of malignant melanoma with immunotherapy is a promising approach to treat advanced stages of the disease. However, the treatment can cause serious side effects and not every patient responds to it. This means, crucial time may be wasted on an ineffective treatment. Assessment of the possible therapy response is hence an important research issue. The research presented in this study focuses on the investigation of the potential of medical imaging and machine learning to solve this task. To this end, we extracted image features from multi-modal images and trained a classifier to differentiate non-responsive patients from responsive ones.
{"title":"Feature-based Response Prediction to Immunotherapy of late-stage Melanoma Patients Using PET/MR Imaging","authors":"Annika Liebgott, S. Gatidis, V. Vu, Tobias Haueise, K. Nikolaou, Bin Yang","doi":"10.23919/Eusipco47968.2020.9287571","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287571","url":null,"abstract":"The treatment of malignant melanoma with immunotherapy is a promising approach to treat advanced stages of the disease. However, the treatment can cause serious side effects and not every patient responds to it. This means, crucial time may be wasted on an ineffective treatment. Assessment of the possible therapy response is hence an important research issue. The research presented in this study focuses on the investigation of the potential of medical imaging and machine learning to solve this task. To this end, we extracted image features from multi-modal images and trained a classifier to differentiate non-responsive patients from responsive ones.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"54 1","pages":"1229-1233"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88116257","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/Eusipco47968.2020.9287216
V. Lostanlen, Alice Cohen-Hadria, J. Bello
With the aim of constructing a biologically plausible model of machine listening, we study the representation of a multicomponent stationary signal by a wavelet scattering network. First, we show that renormalizing second-order nodes by their first-order parents gives a simple numerical criterion to assess whether two neighboring components will interfere psychoacoustically. Secondly, we run a manifold learning algorithm (Isomap) on scattering coefficients to visualize the similarity space underlying parametric additive synthesis. Thirdly, we generalize the “one or two components” framework to three sine waves or more, and prove that the effective scattering depth of a Fourier series grows in logarithmic proportion to its bandwidth.
{"title":"One or Two Frequencies? The Scattering Transform Answers","authors":"V. Lostanlen, Alice Cohen-Hadria, J. Bello","doi":"10.23919/Eusipco47968.2020.9287216","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287216","url":null,"abstract":"With the aim of constructing a biologically plausible model of machine listening, we study the representation of a multicomponent stationary signal by a wavelet scattering network. First, we show that renormalizing second-order nodes by their first-order parents gives a simple numerical criterion to assess whether two neighboring components will interfere psychoacoustically. Secondly, we run a manifold learning algorithm (Isomap) on scattering coefficients to visualize the similarity space underlying parametric additive synthesis. Thirdly, we generalize the “one or two components” framework to three sine waves or more, and prove that the effective scattering depth of a Fourier series grows in logarithmic proportion to its bandwidth.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"11 1","pages":"2205-2209"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87877775","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/Eusipco47968.2020.9287311
Konstantinos Bountrogiannis, G. Tzagkarakis, P. Tsakalides
The ever-increasing volume and complexity of time series data, emerging in various application domains, necessitate efficient dimensionality reduction for facilitating data mining tasks. Symbolic representations, among them symbolic aggregate approximation (SAX), have proven very effective in compacting the information content of time series while exploiting the wealth of search algorithms used in bioinformatics and text mining communities. However, typical SAX-based techniques rely on a Gaussian assumption for the underlying data statistics, which often deteriorates their performance in practical scenarios. To overcome this limitation, this work introduces a method that negates any assumption on the probability distribution of time series. Specifically, a data-driven kernel density estimator is first applied on the data, followed by Lloyd-Max quantization to determine the optimal horizontal segmentation breakpoints. Experimental evaluation on distinct datasets demonstrates the superiority of our method, in terms of reconstruction accuracy and tightness of lower bound, when compared against the conventional and a modified SAX method.
{"title":"Data-driven Kernel-based Probabilistic SAX for Time Series Dimensionality Reduction","authors":"Konstantinos Bountrogiannis, G. Tzagkarakis, P. Tsakalides","doi":"10.23919/Eusipco47968.2020.9287311","DOIUrl":"https://doi.org/10.23919/Eusipco47968.2020.9287311","url":null,"abstract":"The ever-increasing volume and complexity of time series data, emerging in various application domains, necessitate efficient dimensionality reduction for facilitating data mining tasks. Symbolic representations, among them symbolic aggregate approximation (SAX), have proven very effective in compacting the information content of time series while exploiting the wealth of search algorithms used in bioinformatics and text mining communities. However, typical SAX-based techniques rely on a Gaussian assumption for the underlying data statistics, which often deteriorates their performance in practical scenarios. To overcome this limitation, this work introduces a method that negates any assumption on the probability distribution of time series. Specifically, a data-driven kernel density estimator is first applied on the data, followed by Lloyd-Max quantization to determine the optimal horizontal segmentation breakpoints. Experimental evaluation on distinct datasets demonstrates the superiority of our method, in terms of reconstruction accuracy and tightness of lower bound, when compared against the conventional and a modified SAX method.","PeriodicalId":6705,"journal":{"name":"2020 28th European Signal Processing Conference (EUSIPCO)","volume":"59 1","pages":"2343-2347"},"PeriodicalIF":0.0,"publicationDate":"2021-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85829075","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：