Pub Date : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602369
Mostafa El Habib Daho, N. Settouti, Mohammed El Amine Lazouni, M. A. Chikh
Classification systems have been widely applied in different fields such as medical diagnosis. Interpretability represents the most important driving force behind the implementation of fuzzy-based classifiers for medical application problems. Neuro-fuzzy classification approaches aim at creating fuzzy classification rules from data. The simplest model is The NEFCLASS; it is able to learn fuzzy rules and fuzzy sets by simple heuristics. In this paper we present a new hybrid learning algorithm for this model using Particle Swarm Optimization PSO for adjusting membership functions parameters. Experiments are performed on the Pima Indian Diabetes dataset available in UCI machine learning repository. The results indicate that the proposed method can work effectively for classifying the diabetes with an acceptable accuracy and transparency.
{"title":"Recognition of diabetes disease using a new hybrid learning algorithm for NEFCLASS","authors":"Mostafa El Habib Daho, N. Settouti, Mohammed El Amine Lazouni, M. A. Chikh","doi":"10.1109/WOSSPA.2013.6602369","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602369","url":null,"abstract":"Classification systems have been widely applied in different fields such as medical diagnosis. Interpretability represents the most important driving force behind the implementation of fuzzy-based classifiers for medical application problems. Neuro-fuzzy classification approaches aim at creating fuzzy classification rules from data. The simplest model is The NEFCLASS; it is able to learn fuzzy rules and fuzzy sets by simple heuristics. In this paper we present a new hybrid learning algorithm for this model using Particle Swarm Optimization PSO for adjusting membership functions parameters. Experiments are performed on the Pima Indian Diabetes dataset available in UCI machine learning repository. The results indicate that the proposed method can work effectively for classifying the diabetes with an acceptable accuracy and transparency.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"160 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124485952","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602419
S. Dehouche, K. Benachenhou
The American Global Positioning System GPS had indeed monopolized the world of satellite navigation for several years; however, its military use, its limited accuracy and the voluntary degradation of the signal had pushed other countries to develop their own systems. The largest project of its kind in this field, would be, the European Galileo system. The aim of this work is the statistical modeling of the acquisition stage for the signal recently proposed for the European system Galileo AltBoc E5. In this context, a brief study of AltBoc signal is given, then, different architectures acquisition will be modeled and analyzed in a statistical framework with a fixe threshold. Then we propose an introduction of a Cell Averaging Constant False Alarm Rate (CA-CFAR) detector. The obtained theoretical results are validated by Monte-Carlo simulations.
{"title":"Acquisition of the Galileo AltBoc signal with a fixed and adaptive threshold","authors":"S. Dehouche, K. Benachenhou","doi":"10.1109/WOSSPA.2013.6602419","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602419","url":null,"abstract":"The American Global Positioning System GPS had indeed monopolized the world of satellite navigation for several years; however, its military use, its limited accuracy and the voluntary degradation of the signal had pushed other countries to develop their own systems. The largest project of its kind in this field, would be, the European Galileo system. The aim of this work is the statistical modeling of the acquisition stage for the signal recently proposed for the European system Galileo AltBoc E5. In this context, a brief study of AltBoc signal is given, then, different architectures acquisition will be modeled and analyzed in a statistical framework with a fixe threshold. Then we propose an introduction of a Cell Averaging Constant False Alarm Rate (CA-CFAR) detector. The obtained theoretical results are validated by Monte-Carlo simulations.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133919166","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602349
Fouad Bousetouane, C. Motamed, Lynda Dib
Particle Filtering algorithm for tracking the location of an object using a color distribution is one of the most used algorithm in many sub-field of visual tracking problem. However, the use of a color distribution for tracked object description is insufficient in practice. In this paper, we present an adaptive contextual particle filtering algorithm integrating multiple cues to non-rigid object tracking, designed to handle illumination variation, scale change and complex non-rigid motion. For this purpose, low-level contextual information computed through Haralick texture features and color cues are combined into a model describing the appearance of the target. The likelihood of each cue is calculated and the algorithm rely on likelihood factorization as a product of the likelihoods of the cues. Moving object extraction is performed at each frame for initializing the filter and adapting the search space of each particle with the real dimension of the tracked target. Experimental results of applying this approach show improvement in tracking and robustness in recovering from very complex conditions.
{"title":"Contextual adaptive particle filtering for robust real-time non-rigid object tracking","authors":"Fouad Bousetouane, C. Motamed, Lynda Dib","doi":"10.1109/WOSSPA.2013.6602349","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602349","url":null,"abstract":"Particle Filtering algorithm for tracking the location of an object using a color distribution is one of the most used algorithm in many sub-field of visual tracking problem. However, the use of a color distribution for tracked object description is insufficient in practice. In this paper, we present an adaptive contextual particle filtering algorithm integrating multiple cues to non-rigid object tracking, designed to handle illumination variation, scale change and complex non-rigid motion. For this purpose, low-level contextual information computed through Haralick texture features and color cues are combined into a model describing the appearance of the target. The likelihood of each cue is calculated and the algorithm rely on likelihood factorization as a product of the likelihoods of the cues. Moving object extraction is performed at each frame for initializing the filter and adapting the search space of each particle with the real dimension of the tracked target. Experimental results of applying this approach show improvement in tracking and robustness in recovering from very complex conditions.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131149295","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602354
Seif Eddine Azoug, S. Bouguezel
In this paper, we propose a double image encryption method based on the reciprocal-orthogonal parametric (Rap) transform and chaotic maps. In this method, a complex-valued image is constructed by two secret real-valued images, one as amplitude and the other as phase. In addition, two chaotic random phase masks are generated using two non-independent chaotic maps; one mask is multiplied by the resulting complex-valued image before applying the two-dimensional (2-D) Rap transform and the other one is multiplied by the resulting matrix in the transform domain. This step is then followed by a chaotic scrambling between the real and imaginary parts before applying another 2-D Rap transform, which yields the encrypted image. The independent parameters of the Rap transforms and the parameters of the chaotic maps used for the masks and scrambling are successfully exploited as an encryption secret key. Simulation results demonstrate the robustness of the proposed method against blind decryption, brute force and statistical attacks.
{"title":"Double image encryption based on the reciprocal-orthogonal parametric transform and chaotic maps","authors":"Seif Eddine Azoug, S. Bouguezel","doi":"10.1109/WOSSPA.2013.6602354","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602354","url":null,"abstract":"In this paper, we propose a double image encryption method based on the reciprocal-orthogonal parametric (Rap) transform and chaotic maps. In this method, a complex-valued image is constructed by two secret real-valued images, one as amplitude and the other as phase. In addition, two chaotic random phase masks are generated using two non-independent chaotic maps; one mask is multiplied by the resulting complex-valued image before applying the two-dimensional (2-D) Rap transform and the other one is multiplied by the resulting matrix in the transform domain. This step is then followed by a chaotic scrambling between the real and imaginary parts before applying another 2-D Rap transform, which yields the encrypted image. The independent parameters of the Rap transforms and the parameters of the chaotic maps used for the masks and scrambling are successfully exploited as an encryption secret key. Simulation results demonstrate the robustness of the proposed method against blind decryption, brute force and statistical attacks.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124554809","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602393
M. Neggazi, Messaoud Bengherabi, Z. Boulkenafet, A. Amira
This work aims to propose an efficient hardware/software system fo guassian mixture model (GMM) parts-based topology modeling for face identification and verification. Following its great success in speaker recognition, The GMM approach was extended to face recognition providing a good trade-off in terms of complexity, performance and robustness. Despite its reduced complexity compared to other statistical modeling techniques like hiden markov model (HMM) and its variants. The GMM scoring module still to be computationally intensive algorithm consisting of a series of complex tasks executed in sequential order. This constraint limits its suitability for real-time pattern recognition embedded applications. This paper presents an efficient hardware implementation of embedded GMM based classifier. Reconfigurable system in the form of field programmable gate arrays (FPGA) is deployed to embed the hardware part of the proposed system. Furthermore a design of exponential calculation circuit is proposed for the best compromise between effectiveness and complexity. Approximations are also developed to reduce the hardware complexity. The developed system performs the identification process of an unknown input pattern over 200 models in 2.3 seconds, our performance evaluation indicates that a speedup of around S.IX can be achieved over an optimized software implementation running on a 3.3GHz core i3 processor. A results precision of 10-2 is obtained after performing the GMM calculation using the proposed hardware/software system.
{"title":"An efficient FPGA implementation of Gaussian mixture models based classifier: Application to face recognition","authors":"M. Neggazi, Messaoud Bengherabi, Z. Boulkenafet, A. Amira","doi":"10.1109/WOSSPA.2013.6602393","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602393","url":null,"abstract":"This work aims to propose an efficient hardware/software system fo guassian mixture model (GMM) parts-based topology modeling for face identification and verification. Following its great success in speaker recognition, The GMM approach was extended to face recognition providing a good trade-off in terms of complexity, performance and robustness. Despite its reduced complexity compared to other statistical modeling techniques like hiden markov model (HMM) and its variants. The GMM scoring module still to be computationally intensive algorithm consisting of a series of complex tasks executed in sequential order. This constraint limits its suitability for real-time pattern recognition embedded applications. This paper presents an efficient hardware implementation of embedded GMM based classifier. Reconfigurable system in the form of field programmable gate arrays (FPGA) is deployed to embed the hardware part of the proposed system. Furthermore a design of exponential calculation circuit is proposed for the best compromise between effectiveness and complexity. Approximations are also developed to reduce the hardware complexity. The developed system performs the identification process of an unknown input pattern over 200 models in 2.3 seconds, our performance evaluation indicates that a speedup of around S.IX can be achieved over an optimized software implementation running on a 3.3GHz core i3 processor. A results precision of 10-2 is obtained after performing the GMM calculation using the proposed hardware/software system.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"135 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116597805","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602346
M. Khider, B. Haddad, Abdelmalik Taleb Ahmed
This work is based on the use of the theory of large deviations to calculate the grain multifractal spectrum and classify bone micro architecture texture, to do this the multifractal spectrum mode is used, it gives the fractal dimension of the predominant fractal set to detect osteoporosis. In fact, one of the most relevant parameters to differentiate between pathological and normal cases in the trabecular ROI texture is the distance of separation between trabeculae in bone micro architecture. The method we propose here is based on the multifractal analysis of the signal formed by the succession of bone trabecular thickness and trabecular separation obtained from gray level intensities in the trabecular bone texture to classify the two cases of study.
{"title":"Multifractal analysis by the large deviation spectrum to detect osteoporosis","authors":"M. Khider, B. Haddad, Abdelmalik Taleb Ahmed","doi":"10.1109/WOSSPA.2013.6602346","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602346","url":null,"abstract":"This work is based on the use of the theory of large deviations to calculate the grain multifractal spectrum and classify bone micro architecture texture, to do this the multifractal spectrum mode is used, it gives the fractal dimension of the predominant fractal set to detect osteoporosis. In fact, one of the most relevant parameters to differentiate between pathological and normal cases in the trabecular ROI texture is the distance of separation between trabeculae in bone micro architecture. The method we propose here is based on the multifractal analysis of the signal formed by the succession of bone trabecular thickness and trabecular separation obtained from gray level intensities in the trabecular bone texture to classify the two cases of study.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125201493","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602363
Mohamed ElAmine Hadj-Youcef, M. Adnane, A. Bousbia-Salah
In this paper the problematic of epileptic detection is treated. An algorithm of EEG signal classification into two classes: Healthy and Epileptics is developed. The difference with conventional methods is the use of free seizure epileptic records. A good classification accuracy means that it is possible to detect an epileptic in normal state or at an early stage of epilepsy. The raw EEG signal is decomposed using discrete wavelet transform (DWT). Then, principal component analysis (PCA) allows dimensionality reduction and better representation of the data. Several features are extracted and used in support vector machine (SVM) classifier. Results show satisfactory classification accuracy comparable or better than those reported in literature.
{"title":"Detection of epileptics during seizure free periods","authors":"Mohamed ElAmine Hadj-Youcef, M. Adnane, A. Bousbia-Salah","doi":"10.1109/WOSSPA.2013.6602363","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602363","url":null,"abstract":"In this paper the problematic of epileptic detection is treated. An algorithm of EEG signal classification into two classes: Healthy and Epileptics is developed. The difference with conventional methods is the use of free seizure epileptic records. A good classification accuracy means that it is possible to detect an epileptic in normal state or at an early stage of epilepsy. The raw EEG signal is decomposed using discrete wavelet transform (DWT). Then, principal component analysis (PCA) allows dimensionality reduction and better representation of the data. Several features are extracted and used in support vector machine (SVM) classifier. Results show satisfactory classification accuracy comparable or better than those reported in literature.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121046109","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602370
S. Oudjemia, J. Girault, Nour-eddine Derguini, S. Haddab
In this paper, we propose an approach for Medical image analysis to detect tumors and to distinguish between healthy and pathological tissue that are present in the brain and skin. Our analysis is based on wavelet and multifractal formalism. In this analysis, we calculated the best linear regression interval that gives good parameter values calculated from new multiresolution indicator, called the average wavelet coefficient, derived from the wavelet leaders. Two main contributions are brought up: first, we proposed a method for the estimation of multifractal features. Second, we revealed the potential of multifractal features to characterize tumor brain and skin melanoma. We analyzed, compared our estimator and simulated image against wavelet leaders.
{"title":"Multifractal analysis: Application to medical imaging","authors":"S. Oudjemia, J. Girault, Nour-eddine Derguini, S. Haddab","doi":"10.1109/WOSSPA.2013.6602370","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602370","url":null,"abstract":"In this paper, we propose an approach for Medical image analysis to detect tumors and to distinguish between healthy and pathological tissue that are present in the brain and skin. Our analysis is based on wavelet and multifractal formalism. In this analysis, we calculated the best linear regression interval that gives good parameter values calculated from new multiresolution indicator, called the average wavelet coefficient, derived from the wavelet leaders. Two main contributions are brought up: first, we proposed a method for the estimation of multifractal features. Second, we revealed the potential of multifractal features to characterize tumor brain and skin melanoma. We analyzed, compared our estimator and simulated image against wavelet leaders.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127383194","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602345
N. Khalfa, S. Drissi, R. Ghozi, Meriem Jaidane
The aim of this paper is to assess the ability of the electrodermal activity (EDA) to characterize the performance of few amateur and professional Tunisian runners during an annual semi-marathon. We focus on the start and finish phases of the competition. So, we examine the EDA temporal signature during those key phases. We note that the overall EDA performance tends to be similar for all subjects during the starting phase of the competition. How-ever, the end phase seems to differentiate among them: specifically, we note that with a reference subject (win-ner of the semi-marathon) there is a better management of stress level.
{"title":"Temporal signatures of electrodermal activity for the evaluation of runners' performance: Start and finish phases","authors":"N. Khalfa, S. Drissi, R. Ghozi, Meriem Jaidane","doi":"10.1109/WOSSPA.2013.6602345","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602345","url":null,"abstract":"The aim of this paper is to assess the ability of the electrodermal activity (EDA) to characterize the performance of few amateur and professional Tunisian runners during an annual semi-marathon. We focus on the start and finish phases of the competition. So, we examine the EDA temporal signature during those key phases. We note that the overall EDA performance tends to be similar for all subjects during the starting phase of the competition. How-ever, the end phase seems to differentiate among them: specifically, we note that with a reference subject (win-ner of the semi-marathon) there is a better management of stress level.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"83 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127649211","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 : 2013-05-12DOI: 10.1109/WOSSPA.2013.6602386
M. Abidine, B. Fergani, Laurent Clavier
The class imbalance problem is one of the new problems that emerged in activity recognition and that caused suboptimal classification performance. To deal this problem, we propose an efficient way of choosing the suitable regularization parameter C of the Soft-Support Vector Machines (C-SVM) method to perform automatic recognition of activities in a smart home environment. We also discuss how they differ when not considering the weights in C-SVM formulation using cross validation and how it affects their performance. Then, we compare C-SVM with Conditional Random Fields (CRF) considered as the reference method. Our experimental results carried out on three real world imbalanced datasets show that C-SVM based our proposed criterion is capable of solving this class imbalance problem by improving the class accuracy of activity classification compared to other methods.
{"title":"Importance-weighted the imbalanced data for C-SVM classifier to human activity recognition","authors":"M. Abidine, B. Fergani, Laurent Clavier","doi":"10.1109/WOSSPA.2013.6602386","DOIUrl":"https://doi.org/10.1109/WOSSPA.2013.6602386","url":null,"abstract":"The class imbalance problem is one of the new problems that emerged in activity recognition and that caused suboptimal classification performance. To deal this problem, we propose an efficient way of choosing the suitable regularization parameter C of the Soft-Support Vector Machines (C-SVM) method to perform automatic recognition of activities in a smart home environment. We also discuss how they differ when not considering the weights in C-SVM formulation using cross validation and how it affects their performance. Then, we compare C-SVM with Conditional Random Fields (CRF) considered as the reference method. Our experimental results carried out on three real world imbalanced datasets show that C-SVM based our proposed criterion is capable of solving this class imbalance problem by improving the class accuracy of activity classification compared to other methods.","PeriodicalId":417940,"journal":{"name":"2013 8th International Workshop on Systems, Signal Processing and their Applications (WoSSPA)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2013-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127612744","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：