Pub Date : 2012-07-02DOI: 10.1109/ISSPA.2012.6310643
K. Ogbureke, João P. Cabral, Julie Carson-Berndsen
This paper presents a novel approach to explicit duration modelling for HMM-based speech synthesis. The proposed approach is a two-step process. The first step in this process is state level phone alignment and conversion of phone durations into the number of frames. In the second step, a hidden Markov model (HMM) is trained whereby the observation is the number of frames in each state and the hidden state the phone. Finally, the duration of each state (the number of frames) is generated from the trained HMM. Hidden semi-Markov model (HSMM) is the baseline for explicit duration modelling in HMM-based speech synthesis. Both objective and perceptual evaluation on a held-out test set showed comparable results with a baseline HSMM-based speech synthesis. This duration modelling approach is computationally simpler than HSMM and produces comparable results in terms of the quality of synthetic speech.
{"title":"Explicit duration modelling in HMM-based speech synthesis using continuous hidden Markov Model","authors":"K. Ogbureke, João P. Cabral, Julie Carson-Berndsen","doi":"10.1109/ISSPA.2012.6310643","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310643","url":null,"abstract":"This paper presents a novel approach to explicit duration modelling for HMM-based speech synthesis. The proposed approach is a two-step process. The first step in this process is state level phone alignment and conversion of phone durations into the number of frames. In the second step, a hidden Markov model (HMM) is trained whereby the observation is the number of frames in each state and the hidden state the phone. Finally, the duration of each state (the number of frames) is generated from the trained HMM. Hidden semi-Markov model (HSMM) is the baseline for explicit duration modelling in HMM-based speech synthesis. Both objective and perceptual evaluation on a held-out test set showed comparable results with a baseline HSMM-based speech synthesis. This duration modelling approach is computationally simpler than HSMM and produces comparable results in terms of the quality of synthetic speech.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131990065","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310490
Osama A S Alkishriwo, L. Chaparro
Compressive sensing attempts to simplify the frequency transformation and thresholding steps, commonly done in data compression, into one. Sparseness of the signal, in either time or frequency, is required for the convex optimization in compressive sensing to perform well. Although sparseness of certain signals, in either time or frequency, is guaranteed by the uncertainty principle signals composed of chirps are not however sparse in either domain. In this paper we propose an orthogonal linear-chirp transform, the discrete linear chirp transform (DLCT), to represent any signal in terms of linear chirps, with modulation and dual properties. Using the DLCT the sparseness of the signal in either time or frequency can be assessed, and if not sparse in neither of these domains, the modulation and dual properties of the DLCT provide a way to transform the signal into a sparse signal. The application of the proposed DLCT is in data compression. The transformation is illustrated by using sparse and not sparse test signals as well as actual signals.
{"title":"A discrete linear chirp transform (DLCT) for data compression","authors":"Osama A S Alkishriwo, L. Chaparro","doi":"10.1109/ISSPA.2012.6310490","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310490","url":null,"abstract":"Compressive sensing attempts to simplify the frequency transformation and thresholding steps, commonly done in data compression, into one. Sparseness of the signal, in either time or frequency, is required for the convex optimization in compressive sensing to perform well. Although sparseness of certain signals, in either time or frequency, is guaranteed by the uncertainty principle signals composed of chirps are not however sparse in either domain. In this paper we propose an orthogonal linear-chirp transform, the discrete linear chirp transform (DLCT), to represent any signal in terms of linear chirps, with modulation and dual properties. Using the DLCT the sparseness of the signal in either time or frequency can be assessed, and if not sparse in neither of these domains, the modulation and dual properties of the DLCT provide a way to transform the signal into a sparse signal. The application of the proposed DLCT is in data compression. The transformation is illustrated by using sparse and not sparse test signals as well as actual signals.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"1038 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131513220","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310608
Isaac Nickaein, M. Rahmati, S. S. Ghidary, A. Zohrabi
Distributed Video Coding (DVC) is a new class of video coding techniques with the aim of coding the decentralized video sources. While the Stanford Wyner-Ziv codec is a well-known architecture in DVC literature, one of its main drawbacks is the presence of a feedback channel from the decoder to the encoder. This feedback channel makes the use of the codec impractical in some applications. Since the only application of the feedback channel is in requesting more parity bits from the encoder, it could be omitted if the encoder estimates the required parity bits and sends them at once. In this paper, a new method of bitrate estimation using a neural network trained by a new set of features is proposed. In addition, a Hybrid mode is proposed that reduces computational complexity at the decoder in a conventional Wyner-Ziv codec.
{"title":"Feedback-free and hybrid distributed video coding using neural networks","authors":"Isaac Nickaein, M. Rahmati, S. S. Ghidary, A. Zohrabi","doi":"10.1109/ISSPA.2012.6310608","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310608","url":null,"abstract":"Distributed Video Coding (DVC) is a new class of video coding techniques with the aim of coding the decentralized video sources. While the Stanford Wyner-Ziv codec is a well-known architecture in DVC literature, one of its main drawbacks is the presence of a feedback channel from the decoder to the encoder. This feedback channel makes the use of the codec impractical in some applications. Since the only application of the feedback channel is in requesting more parity bits from the encoder, it could be omitted if the encoder estimates the required parity bits and sends them at once. In this paper, a new method of bitrate estimation using a neural network trained by a new set of features is proposed. In addition, a Hybrid mode is proposed that reduces computational complexity at the decoder in a conventional Wyner-Ziv codec.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"115 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134002834","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310566
S. Miralavi, S. Ghorshi, Aidin Tahaei
A major problem in real-time packet-based communication systems, is misrouted or delayed packet which results in degraded perceived voice quality. If packets are not available on time, the packets are considered lost. The easiest solution in a network terminal receiver is to replace silence for the duration of lost speech segments. In a high quality communication system, to avoid degradation in speech quality due to packet loss, a suitable method or algorithm is needed to replace the missing segments of speech. In this paper, we introduce an adaptive filter for replacement of lost speech segment. In this method Kalman filter as a state-space based method will be used to predict the clean speech signal in presence of additive noise. The evaluation results show that Kalman filter has lower MSE compared to other methods in presence of White Gaussian Noise and background noise.
{"title":"A Kalman filter approach to packet loss replacement in presence of additive noise","authors":"S. Miralavi, S. Ghorshi, Aidin Tahaei","doi":"10.1109/ISSPA.2012.6310566","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310566","url":null,"abstract":"A major problem in real-time packet-based communication systems, is misrouted or delayed packet which results in degraded perceived voice quality. If packets are not available on time, the packets are considered lost. The easiest solution in a network terminal receiver is to replace silence for the duration of lost speech segments. In a high quality communication system, to avoid degradation in speech quality due to packet loss, a suitable method or algorithm is needed to replace the missing segments of speech. In this paper, we introduce an adaptive filter for replacement of lost speech segment. In this method Kalman filter as a state-space based method will be used to predict the clean speech signal in presence of additive noise. The evaluation results show that Kalman filter has lower MSE compared to other methods in presence of White Gaussian Noise and background noise.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"268 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132909200","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310660
C. O’Reilly, R. Plamondon
This communication summarizes the outcome of our research program on the design of a diagnostic system for neuromuscular disorders based on the analysis of human movement using the Kinematic Theory of Rapid Human Movements. Herein, this design problem is split in sub-problems which are then described. The solutions adopted at each design step are explained. As an example of application, typical results obtained so far for the assessment of the most important modifiable risk factors of brain stroke (diabetes, hypertension, hypercholesterolemia, obesity, cardiac problems, and cigarette smoking) are reported by the means of the area under the receiver operating characteristic curve (AUC).
{"title":"Design of a neuromuscular disorders diagnostic system using human movement analysis","authors":"C. O’Reilly, R. Plamondon","doi":"10.1109/ISSPA.2012.6310660","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310660","url":null,"abstract":"This communication summarizes the outcome of our research program on the design of a diagnostic system for neuromuscular disorders based on the analysis of human movement using the Kinematic Theory of Rapid Human Movements. Herein, this design problem is split in sub-problems which are then described. The solutions adopted at each design step are explained. As an example of application, typical results obtained so far for the assessment of the most important modifiable risk factors of brain stroke (diabetes, hypertension, hypercholesterolemia, obesity, cardiac problems, and cigarette smoking) are reported by the means of the area under the receiver operating characteristic curve (AUC).","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"273 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121800623","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310463
A. Sadhu, Bo Hu, S. Narasimhan
Wireless sensing technology has gained significant attention in the field of structural health monitoring (SHM). Various decentralized modal identification methods have been developed employing wireless sensors. However, one of themajor bottlenecks - especially dealing with long-term SHM - is the large volume of transmitted data. To overcome this problem, we present compressed sensing as a data reduction preprocessing tool within the framework of blind source separation. The results of source separation are ultimately used for modal identification of linear structures under ambient vibrations. When used together with sparsifying time-frequency decompositions, we show that accurate modal identification results are possible with high compression ratios. The main novelty in the method proposed here is in the application of compressive sensing for decentralized modal identification of civil structures.
{"title":"Blind source separation towards decentralized modal identification using compressive sampling","authors":"A. Sadhu, Bo Hu, S. Narasimhan","doi":"10.1109/ISSPA.2012.6310463","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310463","url":null,"abstract":"Wireless sensing technology has gained significant attention in the field of structural health monitoring (SHM). Various decentralized modal identification methods have been developed employing wireless sensors. However, one of themajor bottlenecks - especially dealing with long-term SHM - is the large volume of transmitted data. To overcome this problem, we present compressed sensing as a data reduction preprocessing tool within the framework of blind source separation. The results of source separation are ultimately used for modal identification of linear structures under ambient vibrations. When used together with sparsifying time-frequency decompositions, we show that accurate modal identification results are possible with high compression ratios. The main novelty in the method proposed here is in the application of compressive sensing for decentralized modal identification of civil structures.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"79 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121044197","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310669
Hadrina Sh-Hussain, S. Salleh, A. K. Ariff, Osama Alhamdani, T. T. Swee, A. M. Noor, H. Oemar, Khalid Yusoff
Humans are different in many ways: fat or thin, young or old, sick or healthy; they may differ in auscultation sites which may vary according to the patient's anatomy. Emphasis must be placed on the characteristics of heart sound based on its intensity which greatly depends on the location of the stethoscope to its pericardium. Each one of these areas will emphasize certain characteristics components of the heart sound. Grouping of the first heart sound (lub) is called the S1 features while the second heart sound (dub) is called the S2 features, the systolic or diastolic features are important factor to determine the types of murmurs. To this end, studies have been limited to reflect on the development and evaluation methods in order to detect the various components constituting signal of the heart sound at one specific auscultation point. The principle area of interest in this paper is, however placing the stethoscope at the semi lunar valve called aortic as position one and pulmonary as position two which will provide better quality of the S2 sound. The S1 heart sound can be heard more clearly in the atroventricle (AV) where the mitral valve as position three and tricuspid valve as position four. Comparative experiments with respect to MFCC feature, different number of HMM states and different number of gaussian mixtures were investigated to measure the influence of these factors on the classification performance at the four locations of auscultation of the heart sound. Interestingly, a five-state model outperformed the four-state model which was supposed to model the four basic components of the heart sounds. It can be said, a five-state average over all Gaussian mixtures model and at the four locations provide the best overall performance of 90.1% accuracy.
{"title":"Application of multipoint auscultation for heart sound diagnostic system (MAHDS)","authors":"Hadrina Sh-Hussain, S. Salleh, A. K. Ariff, Osama Alhamdani, T. T. Swee, A. M. Noor, H. Oemar, Khalid Yusoff","doi":"10.1109/ISSPA.2012.6310669","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310669","url":null,"abstract":"Humans are different in many ways: fat or thin, young or old, sick or healthy; they may differ in auscultation sites which may vary according to the patient's anatomy. Emphasis must be placed on the characteristics of heart sound based on its intensity which greatly depends on the location of the stethoscope to its pericardium. Each one of these areas will emphasize certain characteristics components of the heart sound. Grouping of the first heart sound (lub) is called the S1 features while the second heart sound (dub) is called the S2 features, the systolic or diastolic features are important factor to determine the types of murmurs. To this end, studies have been limited to reflect on the development and evaluation methods in order to detect the various components constituting signal of the heart sound at one specific auscultation point. The principle area of interest in this paper is, however placing the stethoscope at the semi lunar valve called aortic as position one and pulmonary as position two which will provide better quality of the S2 sound. The S1 heart sound can be heard more clearly in the atroventricle (AV) where the mitral valve as position three and tricuspid valve as position four. Comparative experiments with respect to MFCC feature, different number of HMM states and different number of gaussian mixtures were investigated to measure the influence of these factors on the classification performance at the four locations of auscultation of the heart sound. Interestingly, a five-state model outperformed the four-state model which was supposed to model the four basic components of the heart sounds. It can be said, a five-state average over all Gaussian mixtures model and at the four locations provide the best overall performance of 90.1% accuracy.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116128789","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310699
A. Sengodan, W. Cockshott
The main challenge of ground penetrating radar (GPR) based land mine detection is to have an accurate image analysis method that is capable of reducing false alarms. However an accurate image relies on having sufficient spatial resolution in the received signal. But because the diameter of an AP mine can be as low as 2cm and many soils have very high attenuations at frequencies above 3GHz, the accurate detection of landmines is accomplished using advanced algorithms. Using image reconstruction and by carrying out the system level analysis of the issues involved with recognition of landmines allows the landmine detection problem to be solved. The SIMCA ('SIMulated Correlation Algorithm') is a novel and accurate landmine detection tool that carries out correlation between a simulated GPR trace and a clutter1 removed original GPR trace. This correlation is performed using the MATLAB® processing environment. The authors tried using convolution and correlation. But in this paper the correlated results are presented because they produced better results. Validation of the results from the algorithm was done by an expert GPR user and 4 other general users who predict the location of landmines. These predicted results are compared with the ground truth data.
{"title":"The SIMCA algorithm for processing ground penetrating radar data and its use in landmine detection","authors":"A. Sengodan, W. Cockshott","doi":"10.1109/ISSPA.2012.6310699","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310699","url":null,"abstract":"The main challenge of ground penetrating radar (GPR) based land mine detection is to have an accurate image analysis method that is capable of reducing false alarms. However an accurate image relies on having sufficient spatial resolution in the received signal. But because the diameter of an AP mine can be as low as 2cm and many soils have very high attenuations at frequencies above 3GHz, the accurate detection of landmines is accomplished using advanced algorithms. Using image reconstruction and by carrying out the system level analysis of the issues involved with recognition of landmines allows the landmine detection problem to be solved. The SIMCA ('SIMulated Correlation Algorithm') is a novel and accurate landmine detection tool that carries out correlation between a simulated GPR trace and a clutter1 removed original GPR trace. This correlation is performed using the MATLAB® processing environment. The authors tried using convolution and correlation. But in this paper the correlated results are presented because they produced better results. Validation of the results from the algorithm was done by an expert GPR user and 4 other general users who predict the location of landmines. These predicted results are compared with the ground truth data.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116512175","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310665
R. Hedjam, M. Cheriet
In this paper we present a new method for hyperspectral band selection problem. The principle is to create a band adjacency graph (BAG) where the nodes represent the bands and the edges represent the similarity weights between the bands. The Markov Clustering Process (abbreviated MCL process) defines a sequence of stochastic matrices by alternation of two operators on the associated affinity matrix to form distinct clusters of high correlated bands. Each cluster is represented by one band and the representative bands will form the new data cube to be used in subsequent processing. The proposed algorithm is tested on a real dataset and compared against state-of-art. The results are promising.
{"title":"Hyperspectral band selection based on graph clustering","authors":"R. Hedjam, M. Cheriet","doi":"10.1109/ISSPA.2012.6310665","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310665","url":null,"abstract":"In this paper we present a new method for hyperspectral band selection problem. The principle is to create a band adjacency graph (BAG) where the nodes represent the bands and the edges represent the similarity weights between the bands. The Markov Clustering Process (abbreviated MCL process) defines a sequence of stochastic matrices by alternation of two operators on the associated affinity matrix to form distinct clusters of high correlated bands. Each cluster is represented by one band and the representative bands will form the new data cube to be used in subsequent processing. The proposed algorithm is tested on a real dataset and compared against state-of-art. The results are promising.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"88 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116931776","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 : 2012-07-02DOI: 10.1109/ISSPA.2012.6310488
Qian Chen, Guyu Hu, Fang-lin Gu, Peng Xiang
The dynamic time warping (DTW) is a classic similarity measure which can handle the time warping issue in similarity computation of time series. And the DTW with constrained warping window is the most common and practical form of DTW. In this paper, the traditional learning method for optimal warping window of DTW is systematically analyzed. Then the time distance to measure the time deviation between two time series is introduced. Finally a new learning method for optimal warping window size based on DTW and time distance is proposed which can improve DTW classification accuracy with little additional computation. Experimental data show that the optimal DTW with best warping window get better classification accuracy when the new learning method is employed. Additionally, the classification accuracy is better than that of ERP and LCSS, and is close to that of TWED.
{"title":"Learning optimal warping window size of DTW for time series classification","authors":"Qian Chen, Guyu Hu, Fang-lin Gu, Peng Xiang","doi":"10.1109/ISSPA.2012.6310488","DOIUrl":"https://doi.org/10.1109/ISSPA.2012.6310488","url":null,"abstract":"The dynamic time warping (DTW) is a classic similarity measure which can handle the time warping issue in similarity computation of time series. And the DTW with constrained warping window is the most common and practical form of DTW. In this paper, the traditional learning method for optimal warping window of DTW is systematically analyzed. Then the time distance to measure the time deviation between two time series is introduced. Finally a new learning method for optimal warping window size based on DTW and time distance is proposed which can improve DTW classification accuracy with little additional computation. Experimental data show that the optimal DTW with best warping window get better classification accuracy when the new learning method is employed. Additionally, the classification accuracy is better than that of ERP and LCSS, and is close to that of TWED.","PeriodicalId":248763,"journal":{"name":"2012 11th International Conference on Information Science, Signal Processing and their Applications (ISSPA)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2012-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121187727","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：