Pub Date : 2015-07-12DOI: 10.1109/IJCNN.2015.7280811
Lawrence Mutimbu, A. Robles-Kelly
This paper presents a method to recover the pixel-wise illuminant colour for scenes lit by multiple lights. Here, we start from the image formation process and pose the illuminant recovery task in hand into an evidence combining setting. To do this, we construct a factor graph making use of the scale space of the input image and a set of illuminant prototypes. The computation of these prototypes is data driven and, hence, our method is devoid of libraries or user input. The use of a factor graph allows for the illuminant estimates at different scales to be recovered making use of a maximum a posteriori (MAP) inference process. Moreover, we render the computation of the probability marginals used here as exact by constructing our factor graph making use of a Delaunay triangulation. We illustrate the utility of our method for pixelwise illuminant colour recovery on two widely available datasets and compare against a number of alternatives. We also show sample colour correction results on real-world images.
{"title":"Factor graphs for pixelwise illuminant estimation","authors":"Lawrence Mutimbu, A. Robles-Kelly","doi":"10.1109/IJCNN.2015.7280811","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280811","url":null,"abstract":"This paper presents a method to recover the pixel-wise illuminant colour for scenes lit by multiple lights. Here, we start from the image formation process and pose the illuminant recovery task in hand into an evidence combining setting. To do this, we construct a factor graph making use of the scale space of the input image and a set of illuminant prototypes. The computation of these prototypes is data driven and, hence, our method is devoid of libraries or user input. The use of a factor graph allows for the illuminant estimates at different scales to be recovered making use of a maximum a posteriori (MAP) inference process. Moreover, we render the computation of the probability marginals used here as exact by constructing our factor graph making use of a Delaunay triangulation. We illustrate the utility of our method for pixelwise illuminant colour recovery on two widely available datasets and compare against a number of alternatives. We also show sample colour correction results on real-world images.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"27 6 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83508658","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280620
Wentao Mao, Jinwan Wang, Liyun Wang
Presently, the data imbalance problems become more pronounced in the applications of machine learning and pattern recognition. However, many traditional machine learning methods suffer from the imbalanced data which are also collected in online sequential manner. To get fast and efficient classification for this special problem, a new online sequential extreme learning machine method with sequential SMOTE strategy is proposed. The key idea of this method is to reduce the randomness while generating virtual minority samples by means of the distribution characteristic of online sequential data. Utilizing online-sequential extreme learning machine as baseline algorithm, this method contains two stages. In offline stage, principal curve is introduced to model the each class's distribution based on which some virtual samples are generated by synthetic minority over-sampling technique(SMOTE). In online stage, each class's membership is determined according to the projection distance of sample to principal curve. With the help of these memberships, the redundant majority samples as well as unreasonable virtual minority samples are all excluded to lighten the imbalance level in online stage. The proposed method is evaluated on four UCI datasets and the real-world air pollutant forecasting dataset. The experimental results show that, the proposed method outperforms the classical ELM, OS-ELM and SMOTE-based OS-ELM in terms of generalization performance and numerical stability.
{"title":"Online sequential classification of imbalanced data by combining extreme learning machine and improved SMOTE algorithm","authors":"Wentao Mao, Jinwan Wang, Liyun Wang","doi":"10.1109/IJCNN.2015.7280620","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280620","url":null,"abstract":"Presently, the data imbalance problems become more pronounced in the applications of machine learning and pattern recognition. However, many traditional machine learning methods suffer from the imbalanced data which are also collected in online sequential manner. To get fast and efficient classification for this special problem, a new online sequential extreme learning machine method with sequential SMOTE strategy is proposed. The key idea of this method is to reduce the randomness while generating virtual minority samples by means of the distribution characteristic of online sequential data. Utilizing online-sequential extreme learning machine as baseline algorithm, this method contains two stages. In offline stage, principal curve is introduced to model the each class's distribution based on which some virtual samples are generated by synthetic minority over-sampling technique(SMOTE). In online stage, each class's membership is determined according to the projection distance of sample to principal curve. With the help of these memberships, the redundant majority samples as well as unreasonable virtual minority samples are all excluded to lighten the imbalance level in online stage. The proposed method is evaluated on four UCI datasets and the real-world air pollutant forecasting dataset. The experimental results show that, the proposed method outperforms the classical ELM, OS-ELM and SMOTE-based OS-ELM in terms of generalization performance and numerical stability.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"18 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84646720","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280329
A. K. Tanwar, E. Crisostomi, P. Ferraro, Marco Raugi, M. Tucci, G. Giunta
In this paper we used clustering algorithms to compare the typical load profiles of different European countries in different day of the weeks. We find out that better results are obtained if the clustering is not performed directly on the data, but on some features extracted from the data. Clustering results can be exploited by energy providers to tailor more attractive time-varying tariffs for their customers. In particular, despite the relevant differences among the several compared countries, we obtained the interesting result of identifying a single feature that is able to distinguish weekdays from holidays and pre-holidays in all the examined countries.
{"title":"Clustering analysis of the electrical load in european countries","authors":"A. K. Tanwar, E. Crisostomi, P. Ferraro, Marco Raugi, M. Tucci, G. Giunta","doi":"10.1109/IJCNN.2015.7280329","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280329","url":null,"abstract":"In this paper we used clustering algorithms to compare the typical load profiles of different European countries in different day of the weeks. We find out that better results are obtained if the clustering is not performed directly on the data, but on some features extracted from the data. Clustering results can be exploited by energy providers to tailor more attractive time-varying tariffs for their customers. In particular, despite the relevant differences among the several compared countries, we obtained the interesting result of identifying a single feature that is able to distinguish weekdays from holidays and pre-holidays in all the examined countries.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"34 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88507658","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280784
J. A. López, M. Saval-Calvo, Andrés Fuster Guilló, J. G. Rodríguez, Sergio Orts
The automated understanding of people activities from video sequences is an open research topic in which the computer vision and pattern recognition areas have made big efforts in recent years. This paper proposes the Self Organizing Activity Description Map (SOADM). It is a novel neural network based on the self-organizing paradigm to classify high level of semantic understanding from video sequences. The neural network is able to deal with the big gap between human trajectories in a scene and the global behaviour associated to them. Specifically, using simple representations of people trajectories as input, the SOADM is able to both represent and classify human behaviours. Additionally, the map is able to preserve the topological information about the scene. Experiments have been carried out using the Shopping Centre dataset of the CAVIAR database taken into account the global behaviour of an individual. Results confirm the high accuracy of the proposal outperforming previous methods.
{"title":"Self-Organizing Activity Description Map to represent and classify human behaviour","authors":"J. A. López, M. Saval-Calvo, Andrés Fuster Guilló, J. G. Rodríguez, Sergio Orts","doi":"10.1109/IJCNN.2015.7280784","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280784","url":null,"abstract":"The automated understanding of people activities from video sequences is an open research topic in which the computer vision and pattern recognition areas have made big efforts in recent years. This paper proposes the Self Organizing Activity Description Map (SOADM). It is a novel neural network based on the self-organizing paradigm to classify high level of semantic understanding from video sequences. The neural network is able to deal with the big gap between human trajectories in a scene and the global behaviour associated to them. Specifically, using simple representations of people trajectories as input, the SOADM is able to both represent and classify human behaviours. Additionally, the map is able to preserve the topological information about the scene. Experiments have been carried out using the Shopping Centre dataset of the CAVIAR database taken into account the global behaviour of an individual. Results confirm the high accuracy of the proposal outperforming previous methods.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"27 1","pages":"1-7"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88524256","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280810
Jeonghyun Baek, Jisu Kim, Junhyuk Hyun, Euntai Kim
For intelligent vehicle applications, detecting pedestrian technique must be robust and perform in real time. In pedestrian detection, support vector machine (SVM) is one of the popular classifiers because of its robust performance. In this paper, we propose the new method to implement cascade SVM that enables fast rejection of negative samples. The proposed method is tested with INRIA person dataset and show better rejection performance of negative samples than conventional method.
{"title":"New efficient speed-up scheme for cascade implementation of SVM classifier","authors":"Jeonghyun Baek, Jisu Kim, Junhyuk Hyun, Euntai Kim","doi":"10.1109/IJCNN.2015.7280810","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280810","url":null,"abstract":"For intelligent vehicle applications, detecting pedestrian technique must be robust and perform in real time. In pedestrian detection, support vector machine (SVM) is one of the popular classifiers because of its robust performance. In this paper, we propose the new method to implement cascade SVM that enables fast rejection of negative samples. The proposed method is tested with INRIA person dataset and show better rejection performance of negative samples than conventional method.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"37 1","pages":"1-6"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88615242","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280560
B. K. Hedayati, Guangyuan Guangyuan, A. Jooya, N. Dimopoulos
In this paper, the effects of regularization on the generalization capabilities of a neural network model are analyzed. We compare the performance of Levenberg-Marquardt and Bayesian Regularization algorithms with and without post-training regularization. We show that although Bayesian Regularization performs slightly better than Levenberg-Marquardt, the model trained using Levenberg-Marquardt holds more information about the data set which by proper post-processing regularization can be extracted. This post-processing regularization imposes smoothness and similarity.
{"title":"In-training and post-training generalization methods: The case of ppar — α and ppar — γ agonists","authors":"B. K. Hedayati, Guangyuan Guangyuan, A. Jooya, N. Dimopoulos","doi":"10.1109/IJCNN.2015.7280560","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280560","url":null,"abstract":"In this paper, the effects of regularization on the generalization capabilities of a neural network model are analyzed. We compare the performance of Levenberg-Marquardt and Bayesian Regularization algorithms with and without post-training regularization. We show that although Bayesian Regularization performs slightly better than Levenberg-Marquardt, the model trained using Levenberg-Marquardt holds more information about the data set which by proper post-processing regularization can be extracted. This post-processing regularization imposes smoothness and similarity.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"28 1","pages":"1-7"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"89363852","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280689
S. Talebi, S. Kanna, D. Mandic
Frequency estimation in three-phase power systems is considered from a state space point of view, and a robust and fast converging algorithm for estimating the fundamental frequency of three-phase power systems is introduced. This is achieved by exploiting the Clarke transform to incorporate the information from all the phases and then designing a widely linear state space estimator that can accurately estimate the fundamental frequency of both balanced and unbalanced three-phase power systems. The framework is then expanded to modify the state space model in order to account for the presence of harmonics in the system. The performance of the developed algorithm is validated through simulations on both synthetic data and real-world data recordings, where it is shown that the developed algorithm outperforms standard linear and the recently introduced widely liner frequency estimators.
{"title":"A non-linear state space frequency estimator for three-phase power systems","authors":"S. Talebi, S. Kanna, D. Mandic","doi":"10.1109/IJCNN.2015.7280689","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280689","url":null,"abstract":"Frequency estimation in three-phase power systems is considered from a state space point of view, and a robust and fast converging algorithm for estimating the fundamental frequency of three-phase power systems is introduced. This is achieved by exploiting the Clarke transform to incorporate the information from all the phases and then designing a widely linear state space estimator that can accurately estimate the fundamental frequency of both balanced and unbalanced three-phase power systems. The framework is then expanded to modify the state space model in order to account for the presence of harmonics in the system. The performance of the developed algorithm is validated through simulations on both synthetic data and real-world data recordings, where it is shown that the developed algorithm outperforms standard linear and the recently introduced widely liner frequency estimators.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"55 1","pages":"1-7"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84712729","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280604
Wei Jiang, Hao Zheng, Shuai Nie, Wenju Liu
A growing number of noise reduction algorithms based on supervised learning have begun to emerge in recent years and show great promise. In this study, we focus on the problem of speech denoising at very low signal-to-noise ratio (SNR) conditions using artificial neural networks. The overall objective is to increase speech intelligibility in the presence of noise. Inspired by multitask learning (MTL), a novel framework based on deep stacking network (DSN) is proposed to do speech denoising at three different time-frequency scales simultaneously and collaboratively. Experiment results show that our algorithm outperforms a state-of-the-art method that is based on traditional deep neural network (DNN).
{"title":"Multiscale collaborative speech denoising based on deep stacking network","authors":"Wei Jiang, Hao Zheng, Shuai Nie, Wenju Liu","doi":"10.1109/IJCNN.2015.7280604","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280604","url":null,"abstract":"A growing number of noise reduction algorithms based on supervised learning have begun to emerge in recent years and show great promise. In this study, we focus on the problem of speech denoising at very low signal-to-noise ratio (SNR) conditions using artificial neural networks. The overall objective is to increase speech intelligibility in the presence of noise. Inspired by multitask learning (MTL), a novel framework based on deep stacking network (DSN) is proposed to do speech denoising at three different time-frequency scales simultaneously and collaboratively. Experiment results show that our algorithm outperforms a state-of-the-art method that is based on traditional deep neural network (DNN).","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"60 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"84720528","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280770
M. Brunato, R. Battiti
This paper investigates Stochastic Local Search (SLS) algorithms for training neural networks with threshold activation functions. and proposes a novel technique, called Binary Learning Machine (BLM). BLM acts by changing individual bits in the binary representation of each weight and picking improving moves.
{"title":"Stochastic Local Search for direct training of threshold networks","authors":"M. Brunato, R. Battiti","doi":"10.1109/IJCNN.2015.7280770","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280770","url":null,"abstract":"This paper investigates Stochastic Local Search (SLS) algorithms for training neural networks with threshold activation functions. and proposes a novel technique, called Binary Learning Machine (BLM). BLM acts by changing individual bits in the binary representation of each weight and picking improving moves.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"55 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90570877","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 : 2015-07-12DOI: 10.1109/IJCNN.2015.7280838
J. Weng
It has been shown that a Developmental Network (DN) can learn any Finite Automaton (FA) [29] but FA is not a general purpose automaton by itself. This theoretical paper presents that the controller of any Turing Machine (TM) is equivalent to an FA. It further models a motivation-free brain - excluding motivation e.g., emotions - as a TM inside a grounded DN - DN with the real world. Unlike a traditional TM, the TM-in-DN uses natural encoding of input and output and uses emergent internal representations. In Artificial Intelligence (AI) there are two major schools, symbolism and connectionism. The theoretical result here implies that the connectionist school is at least as powerful as the symbolic school also in terms of the general-purpose nature of TM. Furthermore, any TM simulated by the DN is grounded and uses natural encoding so that the DN autonomously learns any TM directly from natural world without a need for a human to encode its input and output. This opens the door for the DN to fully autonomously learn any TM, from a human teacher, reading a book, or real world events. The motivated version of DN [31] further enables a DN to go beyond action-supervised learning - so as to learn based on pain-avoidance, pleasure seeking, and novelty seeking [31].
{"title":"Brains as naturally emerging turing machines","authors":"J. Weng","doi":"10.1109/IJCNN.2015.7280838","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280838","url":null,"abstract":"It has been shown that a Developmental Network (DN) can learn any Finite Automaton (FA) [29] but FA is not a general purpose automaton by itself. This theoretical paper presents that the controller of any Turing Machine (TM) is equivalent to an FA. It further models a motivation-free brain - excluding motivation e.g., emotions - as a TM inside a grounded DN - DN with the real world. Unlike a traditional TM, the TM-in-DN uses natural encoding of input and output and uses emergent internal representations. In Artificial Intelligence (AI) there are two major schools, symbolism and connectionism. The theoretical result here implies that the connectionist school is at least as powerful as the symbolic school also in terms of the general-purpose nature of TM. Furthermore, any TM simulated by the DN is grounded and uses natural encoding so that the DN autonomously learns any TM directly from natural world without a need for a human to encode its input and output. This opens the door for the DN to fully autonomously learn any TM, from a human teacher, reading a book, or real world events. The motivated version of DN [31] further enables a DN to go beyond action-supervised learning - so as to learn based on pain-avoidance, pleasure seeking, and novelty seeking [31].","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"11 1","pages":"1-8"},"PeriodicalIF":0.0,"publicationDate":"2015-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85333440","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：