Pub Date : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280152
A. Mahmood, Hushairi Zen, A. Othman, S. A. Siddiqui
Endowing users of multi-interface mobile handsets the competence to seamlessly roam among diverse heterogeneous wireless networks has become a crucial challenge confronting the network operators / service providers in the recent years. Today, we have moved far beyond the 3G communication networks, wherein, potential to handover traditionally relied on the channel quality computed from the received signal strength (RSS) and the accessibility of resources in new cells. These traditional handover protocols have been envisaged for the homogeneous systems possessing common routing mechanisms, signaling protocols and mobility management standards without taking users' desirable network choice into consideration. On contrary, in heterogeneous environments, mobile devices and network routers must be able to perform `quick handovers of data sessions' amongst different networks and protocols with least possible switching delays and minimized latency, and thus regards the users' desirable network choice as significant. In this manuscript, an optimized handover decision mechanism, `Travelling Time Prediction Based on the Consecutive RSS Measurements' in an IEEE 802.11 WLAN cell is suggested. The method uses a time threshold which is computed by a Mobile Terminal (MT) as soon as it penetrates into a WLAN boundary. The estimated travelling time is then compared with the time threshold, so as to make handover decisions for reducing the probability of handover failures. Our performance analysis reveals that the suggested mechanism effectively minimizes the number of handover failures by 60% as compared to the already proposed schemes.
{"title":"An optimized travelling time estimation mechanism for minimizing handover failures from cellular networks to WLANs","authors":"A. Mahmood, Hushairi Zen, A. Othman, S. A. Siddiqui","doi":"10.1109/ICEDIF.2015.7280152","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280152","url":null,"abstract":"Endowing users of multi-interface mobile handsets the competence to seamlessly roam among diverse heterogeneous wireless networks has become a crucial challenge confronting the network operators / service providers in the recent years. Today, we have moved far beyond the 3G communication networks, wherein, potential to handover traditionally relied on the channel quality computed from the received signal strength (RSS) and the accessibility of resources in new cells. These traditional handover protocols have been envisaged for the homogeneous systems possessing common routing mechanisms, signaling protocols and mobility management standards without taking users' desirable network choice into consideration. On contrary, in heterogeneous environments, mobile devices and network routers must be able to perform `quick handovers of data sessions' amongst different networks and protocols with least possible switching delays and minimized latency, and thus regards the users' desirable network choice as significant. In this manuscript, an optimized handover decision mechanism, `Travelling Time Prediction Based on the Consecutive RSS Measurements' in an IEEE 802.11 WLAN cell is suggested. The method uses a time threshold which is computed by a Mobile Terminal (MT) as soon as it penetrates into a WLAN boundary. The estimated travelling time is then compared with the time threshold, so as to make handover decisions for reducing the probability of handover failures. Our performance analysis reveals that the suggested mechanism effectively minimizes the number of handover failures by 60% as compared to the already proposed schemes.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115356556","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280200
Dong Zhu, Linji Lu
Order tracking technique is an effective frequency analysis method, which uses multiples of the running speed as the frequency base (orders) and commonly used in rotating machinery vibration signal analysis. It is a dedicated non-stationary vibration processing technique to detect speed-related vibrations. Angular sampling theory based computed order tracking (COT) method is the most widely used method of order tracking. It samples the vibration at a constant rate, and then uses software to resample the sampled data at constant angle increments. Previous research indicates that the computed order components obtained through COT method are not precise, and the program is of high computation complexity. To make COT more accurate and efficient, this paper presents an improved resampling method for COT, which is inspired by time-frequency scaling property of Fourier Transform (FT). A stretching transformation is used to stretch the sampled data on time domain by rotating speed before the resampling process, and the product of the stretching transformation is proved to be non-constant samples of order domain. Simulated results demonstrate the improvements of the proposed method on order tracking accuracy and a lower computation complexity, especially for signals with high order components. At last, a turbocharger quality inspection system is designed, in which the presented method is used to analyze vibrations generated by the turbocharger running under a controlled running-up condition. The testing result indicates that the proposed method works well on extracting harmonic vibrations from the sampled data.
{"title":"Resampling method of computed order tracking based on time-frequency scaling property of fourier transform","authors":"Dong Zhu, Linji Lu","doi":"10.1109/ICEDIF.2015.7280200","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280200","url":null,"abstract":"Order tracking technique is an effective frequency analysis method, which uses multiples of the running speed as the frequency base (orders) and commonly used in rotating machinery vibration signal analysis. It is a dedicated non-stationary vibration processing technique to detect speed-related vibrations. Angular sampling theory based computed order tracking (COT) method is the most widely used method of order tracking. It samples the vibration at a constant rate, and then uses software to resample the sampled data at constant angle increments. Previous research indicates that the computed order components obtained through COT method are not precise, and the program is of high computation complexity. To make COT more accurate and efficient, this paper presents an improved resampling method for COT, which is inspired by time-frequency scaling property of Fourier Transform (FT). A stretching transformation is used to stretch the sampled data on time domain by rotating speed before the resampling process, and the product of the stretching transformation is proved to be non-constant samples of order domain. Simulated results demonstrate the improvements of the proposed method on order tracking accuracy and a lower computation complexity, especially for signals with high order components. At last, a turbocharger quality inspection system is designed, in which the presented method is used to analyze vibrations generated by the turbocharger running under a controlled running-up condition. The testing result indicates that the proposed method works well on extracting harmonic vibrations from the sampled data.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"25 4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123243972","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280165
Wandi Du, Hanshi Wang, Lizhen Liu, Wei Song, Jingli Lu
Fingerprint segmentation is the key step of fingerprint image preprocessing. Efficient fingerprint segmentation technology has significance in both saving preprocessing time and improving the image quality. In this paper, on the basis of the right direction of the fingerprint ridge, we use the gradient threshold method to segment image for the first time. While there are still limitations on the performance of the first segmentation, the second segmentation is used to improve the quality of results, which is based on matrix manipulation. Experimental results prove that this method has better denoising performance and higher computing speed. Finally, we get the high-quality fingerprint image.
{"title":"The optimization of fingerprint segmentation based on sparse representation","authors":"Wandi Du, Hanshi Wang, Lizhen Liu, Wei Song, Jingli Lu","doi":"10.1109/ICEDIF.2015.7280165","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280165","url":null,"abstract":"Fingerprint segmentation is the key step of fingerprint image preprocessing. Efficient fingerprint segmentation technology has significance in both saving preprocessing time and improving the image quality. In this paper, on the basis of the right direction of the fingerprint ridge, we use the gradient threshold method to segment image for the first time. While there are still limitations on the performance of the first segmentation, the second segmentation is used to improve the quality of results, which is based on matrix manipulation. Experimental results prove that this method has better denoising performance and higher computing speed. Finally, we get the high-quality fingerprint image.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"97 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125030636","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280171
Jun Wang, Yuan Gao, C. Ran, Yinlong Huo
In order to obtain more accurate state estimation from multisensor system, the state estimation with two-level fusion structure is presented. By means of measurement fusion algorithm, the local first-level fusion centers can obtain the globally optimal fused measurement information, and then the local state estimation can be got by classical Kalman filtering. At the second-level fusion center, the fused estimation can be received by applying the covariance intersection fusion algorithm, which avoids the calculation of the correlation among local first-level fusion centers. The simulation example shows the effectiveness and higher accuracy of the presented fusion structure.
{"title":"State estimation with two-level fusion structure","authors":"Jun Wang, Yuan Gao, C. Ran, Yinlong Huo","doi":"10.1109/ICEDIF.2015.7280171","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280171","url":null,"abstract":"In order to obtain more accurate state estimation from multisensor system, the state estimation with two-level fusion structure is presented. By means of measurement fusion algorithm, the local first-level fusion centers can obtain the globally optimal fused measurement information, and then the local state estimation can be got by classical Kalman filtering. At the second-level fusion center, the fused estimation can be received by applying the covariance intersection fusion algorithm, which avoids the calculation of the correlation among local first-level fusion centers. The simulation example shows the effectiveness and higher accuracy of the presented fusion structure.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130254540","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280173
C. Ran, Y. Dou, Yuan Gao
For the multisensor linear stochastic descriptor system with same measurement matrix and correlated noises, the weighted measurement fusion information filter is presented, based on the weighted measurement fusion algorithm and the Kalman information filtering theory. This information filtering is a new repression of Kalman filtering based on information matrix, which can reduce computational burden and has important application in many theory analysis. And the presented weighted measurement fusion information filter has global optimality, and can avoid computing these cross-variances of the local Kalman filters, compared with the state fusion method. A simulation example about 3-sensors stochastic descriptor system verifies the effectiveness.
{"title":"WMF information filter for multisensor descriptor system with same measurement matrix and correlated noises","authors":"C. Ran, Y. Dou, Yuan Gao","doi":"10.1109/ICEDIF.2015.7280173","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280173","url":null,"abstract":"For the multisensor linear stochastic descriptor system with same measurement matrix and correlated noises, the weighted measurement fusion information filter is presented, based on the weighted measurement fusion algorithm and the Kalman information filtering theory. This information filtering is a new repression of Kalman filtering based on information matrix, which can reduce computational burden and has important application in many theory analysis. And the presented weighted measurement fusion information filter has global optimality, and can avoid computing these cross-variances of the local Kalman filters, compared with the state fusion method. A simulation example about 3-sensors stochastic descriptor system verifies the effectiveness.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125139342","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280150
Hongqing Liu, Yong Li, Yi Zhou, Jianzhong Huang
In this work, we develop a minimum mean square error (MMSE) estimator for the underdetermined systems when the signal of interest is sparse. To address the uncertainty issue introduced in the measurement system, robust approaches are developed based on stochastic and worst case optimization techniques under the minimax framework. To solve the optimization problem, different constraints on the unknown signal of interest are considered to transform the minimax optimization into semidefinite programming problem (SDP), which can be efficiently solved. Numerical studies are provided to demonstrate utilizing sparsity and robust approaches indeed improve MMSE estimator when the sparsity of the signal of interest is utilized and the system considered is underdetermined.
{"title":"Robust minimax MMSE for sparse signal recovery against system perturbations","authors":"Hongqing Liu, Yong Li, Yi Zhou, Jianzhong Huang","doi":"10.1109/ICEDIF.2015.7280150","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280150","url":null,"abstract":"In this work, we develop a minimum mean square error (MMSE) estimator for the underdetermined systems when the signal of interest is sparse. To address the uncertainty issue introduced in the measurement system, robust approaches are developed based on stochastic and worst case optimization techniques under the minimax framework. To solve the optimization problem, different constraints on the unknown signal of interest are considered to transform the minimax optimization into semidefinite programming problem (SDP), which can be efficiently solved. Numerical studies are provided to demonstrate utilizing sparsity and robust approaches indeed improve MMSE estimator when the sparsity of the signal of interest is utilized and the system considered is underdetermined.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121047747","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280167
Guangfu Zhou, Chenglin Wen, Jingli Gao
With the rapid development of modern information technology, target recognition plays an increasingly important role in agricultural production, national defense construction. However, the existing target recognition algorithm has many limitations, such as image distortion, difficult to recognize target image or poor recognition results because of camera angles and lighting conditions. Based on the above questions, the paper proposes an image recognition algorithm, and the light field is applied to the image recognition as the feature extraction library for first time. First, we obtain light field information which contains images taken from different angles of the target object, and then regard the light field information as an object library. Finally we perform the algorithm of target recognition for the target image based on the object library. Based on sparse Fourier transform, the light field reconstruction algorithm in this paper can reconstruct the entire light field with a small amount of samples. This recognition algorithm can solve the problem to recognize image due to the different camera angles. Finally, the simulation verifies the effectiveness of the algorithm.
{"title":"Object recognition based on reconstruction of light field","authors":"Guangfu Zhou, Chenglin Wen, Jingli Gao","doi":"10.1109/ICEDIF.2015.7280167","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280167","url":null,"abstract":"With the rapid development of modern information technology, target recognition plays an increasingly important role in agricultural production, national defense construction. However, the existing target recognition algorithm has many limitations, such as image distortion, difficult to recognize target image or poor recognition results because of camera angles and lighting conditions. Based on the above questions, the paper proposes an image recognition algorithm, and the light field is applied to the image recognition as the feature extraction library for first time. First, we obtain light field information which contains images taken from different angles of the target object, and then regard the light field information as an object library. Finally we perform the algorithm of target recognition for the target image based on the object library. Based on sparse Fourier transform, the light field reconstruction algorithm in this paper can reconstruct the entire light field with a small amount of samples. This recognition algorithm can solve the problem to recognize image due to the different camera angles. Finally, the simulation verifies the effectiveness of the algorithm.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116298431","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280188
Wenqiang Liu, Z. Deng
In this paper, the problem of designing robust steady-state Kalman filter is considered for linear discrete-time system with uncertain model parameters and noise variances. By the new approach of compensating the parameter uncertainties by a fictitious noise, the system model is converted into that with uncertain noise variances only. Using the minimax robust estimation principle, based on the worst-case conservative system with the conservative upper bounds of the noise variances, a robust steady-state Kalman filter is presented. Based on the Lyapunov equation approach, we prove its robustness. The concept of the robust region is presented. A simulation example is presented to demonstrate how to search the robust region and show its good performance.
{"title":"Robust steady-state Kalman filter for uncertain discrete-time system","authors":"Wenqiang Liu, Z. Deng","doi":"10.1109/ICEDIF.2015.7280188","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280188","url":null,"abstract":"In this paper, the problem of designing robust steady-state Kalman filter is considered for linear discrete-time system with uncertain model parameters and noise variances. By the new approach of compensating the parameter uncertainties by a fictitious noise, the system model is converted into that with uncertain noise variances only. Using the minimax robust estimation principle, based on the worst-case conservative system with the conservative upper bounds of the noise variances, a robust steady-state Kalman filter is presented. Based on the Lyapunov equation approach, we prove its robustness. The concept of the robust region is presented. A simulation example is presented to demonstrate how to search the robust region and show its good performance.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123874240","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280185
Hua Juliang, Liang Haicheng, Li Shijin
This paper proposes a fast algorithm for multiple targets tracking in complex environment of industrial workshop, which integrates the background modeling and the motion information. First, the probability density image is calculated based on histogram of color from each target object. Second, these probability density images are filtered according to background image obtained from previous background modeling. Third, the motion information is fused into its tracking process, and the optimal position is thus predicted. Finally, the algorithm removes the false targets in the previous frame from those images of color probability density, in order to avoid the disturbance to other targets in the later tracking procedure. The experimental results have demonstrated that the proposed new algorithm is capable of reducing background and similar objects disturbance and achieving real-time performance.
{"title":"A fast multi-object tracking algorithm by fusing color and motion information","authors":"Hua Juliang, Liang Haicheng, Li Shijin","doi":"10.1109/ICEDIF.2015.7280185","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280185","url":null,"abstract":"This paper proposes a fast algorithm for multiple targets tracking in complex environment of industrial workshop, which integrates the background modeling and the motion information. First, the probability density image is calculated based on histogram of color from each target object. Second, these probability density images are filtered according to background image obtained from previous background modeling. Third, the motion information is fused into its tracking process, and the optimal position is thus predicted. Finally, the algorithm removes the false targets in the previous frame from those images of color probability density, in order to avoid the disturbance to other targets in the later tracking procedure. The experimental results have demonstrated that the proposed new algorithm is capable of reducing background and similar objects disturbance and achieving real-time performance.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130166095","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 : 2015-10-01DOI: 10.1109/ICEDIF.2015.7280166
Dongli Wang, Yanhua Wei, Yan Zhou, Tingrui Pei
Based on fisher ratio class separability measure, we propose two types of posterior probability support vector machines (PPSVMs) using binary tree structure. The first one is a some-against-rest binary tree of PPSVM classifiers (SBT), for which some classes as a cluster are divided from the rest classes at each non-leaf node. To determine the two clusters, we use the Fisher ratio separability measure. Accordingly, the second proposed method termed one-against-rest binary tree of PPSVMs (OBT), we separate only one class with the largest separability measure from the rest classes at each non-leaf node. Then, the procedures of both SBT and OBT are provided. Finally, we consider the problem of human action recognition based on depth maps adopting both proposed approaches. Simulation results indicate both methods gain higher classifying accuracy than those of canonical multi-class SVMs and PPSVMs. Besides, the decision complexity of the proposed SBT and OBT are reduced because they use the posterior probability and the Fisher ratio separability measure.
{"title":"Fisher-ratio-separability boosted binary tree of posterior probability SVMs with application to action recognition","authors":"Dongli Wang, Yanhua Wei, Yan Zhou, Tingrui Pei","doi":"10.1109/ICEDIF.2015.7280166","DOIUrl":"https://doi.org/10.1109/ICEDIF.2015.7280166","url":null,"abstract":"Based on fisher ratio class separability measure, we propose two types of posterior probability support vector machines (PPSVMs) using binary tree structure. The first one is a some-against-rest binary tree of PPSVM classifiers (SBT), for which some classes as a cluster are divided from the rest classes at each non-leaf node. To determine the two clusters, we use the Fisher ratio separability measure. Accordingly, the second proposed method termed one-against-rest binary tree of PPSVMs (OBT), we separate only one class with the largest separability measure from the rest classes at each non-leaf node. Then, the procedures of both SBT and OBT are provided. Finally, we consider the problem of human action recognition based on depth maps adopting both proposed approaches. Simulation results indicate both methods gain higher classifying accuracy than those of canonical multi-class SVMs and PPSVMs. Besides, the decision complexity of the proposed SBT and OBT are reduced because they use the posterior probability and the Fisher ratio separability measure.","PeriodicalId":355975,"journal":{"name":"2015 International Conference on Estimation, Detection and Information Fusion (ICEDIF)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129297703","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：