Pub Date : 2015-07-12DOI: 10.1109/IJCNN.2015.7280639
Andrea Schnall, M. Heckmann
In this paper we investigate optical flow field features for the automatic labeling of word prominence. Visual motion is a rich source of information. Modifying the articulatory parameters to raise the prominence of a segment of an utterance, is usually accompanied by a stronger movement of mouth and head compared to a non-prominent segment. One way to describe such motion is to use optical flow fields. During the recording of the audio-visual database we used for the following experiments, the subjects were asked to make corrections for a misunderstanding of a single word of the system by using prosodic cues only, which created a narrow and a broad focus. Audio-visual recordings with a distant microphone and without visual markers were made. As acoustic features duration, loudness, fundamental frequency and spectral emphasis were calculated. From the visual channel the nose position is detected and the mouth region is extracted. From this region the optical flow is calculated and all the optical flow fields for one word are summed up. The pooled optical flow for the four directions is then used as feature vector. We demonstrate that using these features in addition to the audio features can improve the classification results for some speakers. We also compare the optical flow field features to other visual features, the nose position and image transformation based visual features. The optical flow field features incorporate not as much information as image transformation based visual features, but using both in addition to the audio features leads to the overall best results, which shows that they contain complementary information.
{"title":"Evaluation of optical flow field features for the detection of word prominence in a human-machine interaction scenario","authors":"Andrea Schnall, M. Heckmann","doi":"10.1109/IJCNN.2015.7280639","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280639","url":null,"abstract":"In this paper we investigate optical flow field features for the automatic labeling of word prominence. Visual motion is a rich source of information. Modifying the articulatory parameters to raise the prominence of a segment of an utterance, is usually accompanied by a stronger movement of mouth and head compared to a non-prominent segment. One way to describe such motion is to use optical flow fields. During the recording of the audio-visual database we used for the following experiments, the subjects were asked to make corrections for a misunderstanding of a single word of the system by using prosodic cues only, which created a narrow and a broad focus. Audio-visual recordings with a distant microphone and without visual markers were made. As acoustic features duration, loudness, fundamental frequency and spectral emphasis were calculated. From the visual channel the nose position is detected and the mouth region is extracted. From this region the optical flow is calculated and all the optical flow fields for one word are summed up. The pooled optical flow for the four directions is then used as feature vector. We demonstrate that using these features in addition to the audio features can improve the classification results for some speakers. We also compare the optical flow field features to other visual features, the nose position and image transformation based visual features. The optical flow field features incorporate not as much information as image transformation based visual features, but using both in addition to the audio features leads to the overall best results, which shows that they contain complementary information.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"36 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":"85757232","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.2015.7280532
T. Jacob, W. Snyder
We present a new learning rule for intralayer connections in neural networks. The rule is based on Hebbian learning principles and is derived from information theoretic considerations. A simple network trained using the rule is shown to have associative memory like properties. The network acts by building connections between correlated data points, under constraints.
{"title":"Learning rule for associative memory in recurrent neural networks","authors":"T. Jacob, W. Snyder","doi":"10.1109/IJCNN.2015.7280532","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280532","url":null,"abstract":"We present a new learning rule for intralayer connections in neural networks. The rule is based on Hebbian learning principles and is derived from information theoretic considerations. A simple network trained using the rule is shown to have associative memory like properties. The network acts by building connections between correlated data points, under constraints.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"62 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":"76607550","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.2015.7280793
R. Calvo, A. A. Constantino, M. Figueiredo
A bio-inspired coordination strategy is improved aiming at reducing the time needed to mobile multiagent systems accomplish surveillance tasks. The original strategy is based on a modified version of the basic ant system algorithm. Only repulsive pheromone are considered in that version. The new strategy version uses other two kinds of pheromone, as well. Now the agents are able to mark strategic locations for reducing the path length. At the beginning the agents try any trajectory to accomplish the surveillance task. After many trials the agents choose preferably the paths that reduce the total length (time) needed to complete the surveillance task. Comparisons between the original and the extended coordination strategies are presented. Results show that the extended strategy achieves the best performance, that is, the surveillance tasks are accomplished in a period of time shorter than that needed by the original one in view of the scenarios examined.
{"title":"A multi-pheromone stigmergic distributed robot coordination strategy for fast surveillance task execution in unknown environments","authors":"R. Calvo, A. A. Constantino, M. Figueiredo","doi":"10.1109/IJCNN.2015.7280793","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280793","url":null,"abstract":"A bio-inspired coordination strategy is improved aiming at reducing the time needed to mobile multiagent systems accomplish surveillance tasks. The original strategy is based on a modified version of the basic ant system algorithm. Only repulsive pheromone are considered in that version. The new strategy version uses other two kinds of pheromone, as well. Now the agents are able to mark strategic locations for reducing the path length. At the beginning the agents try any trajectory to accomplish the surveillance task. After many trials the agents choose preferably the paths that reduce the total length (time) needed to complete the surveillance task. Comparisons between the original and the extended coordination strategies are presented. Results show that the extended strategy achieves the best performance, that is, the surveillance tasks are accomplished in a period of time shorter than that needed by the original one in view of the scenarios examined.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"51 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":"85506946","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.2015.7280489
Chao Dong, Bo Zhou, Jinglu Hu
This paper presents a new strategy to build multi tree hierarchical structure SVM which can get a more efficient and accuracy classification model for multiclass problems. Base on the theory of Binary Tree SVM (BTS), we proposed an improvement algorithm which extend binary tree structure to a multi tree structure, In the multi tree hierarchical structure, similarity clustering method was proposed to cluster classes to groups in each non-leaf node. In order to get a multi node division, one-against-all (OAA) was applied to train those groups rather than classes. The proposed method can avoid data imbalanced problem occurred in OAA, also the classification area of classifier in the upper layer is larger than classifier in lower layer. Compared with other several well-known methods, experiments on many data sets demonstrate that our method can reduce the number of classifiers in the testing phase and get a higher accuracy.
{"title":"A hierarchical SVM based multiclass classification by using similarity clustering","authors":"Chao Dong, Bo Zhou, Jinglu Hu","doi":"10.1109/IJCNN.2015.7280489","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280489","url":null,"abstract":"This paper presents a new strategy to build multi tree hierarchical structure SVM which can get a more efficient and accuracy classification model for multiclass problems. Base on the theory of Binary Tree SVM (BTS), we proposed an improvement algorithm which extend binary tree structure to a multi tree structure, In the multi tree hierarchical structure, similarity clustering method was proposed to cluster classes to groups in each non-leaf node. In order to get a multi node division, one-against-all (OAA) was applied to train those groups rather than classes. The proposed method can avoid data imbalanced problem occurred in OAA, also the classification area of classifier in the upper layer is larger than classifier in lower layer. Compared with other several well-known methods, experiments on many data sets demonstrate that our method can reduce the number of classifiers in the testing phase and get a higher accuracy.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"53 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":"77929769","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.2015.7280387
Y. Hayashi, Shota Fujisawa
In this paper, we review all our work since 2012 and propose a strategic approach for the Multiple-MLP Ensemble Re- RX algorithm. We first describe the background and procedures of the Recursive-Rule Extraction (Re-RX) algorithm family and its variants, including the Multiple-MLP Ensemble Re-RX algorithm (“Multiple-MLP Ensemble”), which uses the Re-RX algorithm as its core. The proposed strategic approach consists of two processes: non-pruning for the trained neural network ensembles without continuous attributes and a relaxed rule generation scheme using continuous attributes to extract extremely accurate, comprehensible, and concise rules for multi-class mixed datasets (i.e., discrete attributes and continuous attributes). We conducted experiments to find rules for seven kinds of multi-class mixed datasets and compared the accuracy, comprehensibility, and conciseness for the Multiple-MLP Ensemble Re-RX algorithm. The strategic approach for the Multiple-MLP Ensemble Re-RX algorithm outperformed the original Multiple-MLP Ensemble Re- RX algorithm. These results confirm that the strategic approach for the Multiple-MLP Ensemble algorithm facilitates the migration from existing data systems toward new accurate analytic systems and Big Data.
{"title":"Strategic approach for Multiple-MLP Ensemble Re-RX algorithm","authors":"Y. Hayashi, Shota Fujisawa","doi":"10.1109/IJCNN.2015.7280387","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280387","url":null,"abstract":"In this paper, we review all our work since 2012 and propose a strategic approach for the Multiple-MLP Ensemble Re- RX algorithm. We first describe the background and procedures of the Recursive-Rule Extraction (Re-RX) algorithm family and its variants, including the Multiple-MLP Ensemble Re-RX algorithm (“Multiple-MLP Ensemble”), which uses the Re-RX algorithm as its core. The proposed strategic approach consists of two processes: non-pruning for the trained neural network ensembles without continuous attributes and a relaxed rule generation scheme using continuous attributes to extract extremely accurate, comprehensible, and concise rules for multi-class mixed datasets (i.e., discrete attributes and continuous attributes). We conducted experiments to find rules for seven kinds of multi-class mixed datasets and compared the accuracy, comprehensibility, and conciseness for the Multiple-MLP Ensemble Re-RX algorithm. The strategic approach for the Multiple-MLP Ensemble Re-RX algorithm outperformed the original Multiple-MLP Ensemble Re- RX algorithm. These results confirm that the strategic approach for the Multiple-MLP Ensemble algorithm facilitates the migration from existing data systems toward new accurate analytic systems and Big Data.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"37 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":"78060727","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.2015.7280769
Stephen C. Ashmore, Michael S. Gashler
We present Forward Bipartite Alignment (FBA), a method that aligns the topological structures of two neural networks. Neural networks are considered to be a black box, because neural networks have a complex model surface determined by their weights that combine attributes non-linearly. Two networks that make similar predictions on training data may still generalize differently. FBA enables a diversity of applications, including visualization and canonicalization of neural networks, ensembles, and cross-over between unrelated neural networks in evolutionary optimization. We describe the FBA algorithm, and describe implementations for three applications: genetic algorithms, visualization, and ensembles. We demonstrate FBA's usefulness by comparing a bag of neural networks to a bag of FBA-aligned neural networks. We also show that aligning, and then combining two neural networks has no appreciable loss in accuracy which means that Forward Bipartite Alignment aligns neural networks in a meaningful way.
{"title":"A method for finding similarity between multi-layer perceptrons by Forward Bipartite Alignment","authors":"Stephen C. Ashmore, Michael S. Gashler","doi":"10.1109/IJCNN.2015.7280769","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280769","url":null,"abstract":"We present Forward Bipartite Alignment (FBA), a method that aligns the topological structures of two neural networks. Neural networks are considered to be a black box, because neural networks have a complex model surface determined by their weights that combine attributes non-linearly. Two networks that make similar predictions on training data may still generalize differently. FBA enables a diversity of applications, including visualization and canonicalization of neural networks, ensembles, and cross-over between unrelated neural networks in evolutionary optimization. We describe the FBA algorithm, and describe implementations for three applications: genetic algorithms, visualization, and ensembles. We demonstrate FBA's usefulness by comparing a bag of neural networks to a bag of FBA-aligned neural networks. We also show that aligning, and then combining two neural networks has no appreciable loss in accuracy which means that Forward Bipartite Alignment aligns neural networks in a meaningful way.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"15 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":"78107397","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.2015.7280396
Ying-Wei Tan, Wenju Liu, Wei Jiang, Hao Zheng
Speech production knowledge has been used to enhance the phonetic representation and the performance of automatic speech recognition (ASR) systems successfully. Representations of speech production make simple explanations for many phenomena observed in speech. These phenomena can not be easily analyzed from either acoustic signal or phonetic transcription alone. One of the most important aspects of speech production knowledge is the use of articulatory knowledge, which describes the smooth and continuous movements in the vocal tract. In this paper, we present a new articulatory model to provide available information for rescoring the speech recognition lattice hypothesis. The articulatory model consists of a feature front-end, which computes a voicing feature based on a spectral harmonics correlation (SHC) function, and a back-end based on the combination of deep neural networks (DNNs) and hidden Markov models (HMMs). The voicing features are incorporated with standard Mel frequency cepstral coefficients (MFCCs) using heteroscedastic linear discriminant analysis (HLDA) to compensate the speech recognition accuracy rates. Moreover, the advantages of two different models are taken into account by the algorithm, which retains deep learning properties of DNNs, while modeling the articulatory context powerfully through HMMs. Mandarin speech recognition experiments show the proposed method achieves significant improvements in speech recognition performance over the system using MFCCs alone.
{"title":"Integration of articulatory knowledge and voicing features based on DNN/HMM for Mandarin speech recognition","authors":"Ying-Wei Tan, Wenju Liu, Wei Jiang, Hao Zheng","doi":"10.1109/IJCNN.2015.7280396","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280396","url":null,"abstract":"Speech production knowledge has been used to enhance the phonetic representation and the performance of automatic speech recognition (ASR) systems successfully. Representations of speech production make simple explanations for many phenomena observed in speech. These phenomena can not be easily analyzed from either acoustic signal or phonetic transcription alone. One of the most important aspects of speech production knowledge is the use of articulatory knowledge, which describes the smooth and continuous movements in the vocal tract. In this paper, we present a new articulatory model to provide available information for rescoring the speech recognition lattice hypothesis. The articulatory model consists of a feature front-end, which computes a voicing feature based on a spectral harmonics correlation (SHC) function, and a back-end based on the combination of deep neural networks (DNNs) and hidden Markov models (HMMs). The voicing features are incorporated with standard Mel frequency cepstral coefficients (MFCCs) using heteroscedastic linear discriminant analysis (HLDA) to compensate the speech recognition accuracy rates. Moreover, the advantages of two different models are taken into account by the algorithm, which retains deep learning properties of DNNs, while modeling the articulatory context powerfully through HMMs. Mandarin speech recognition experiments show the proposed method achieves significant improvements in speech recognition performance over the system using MFCCs alone.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"50 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":"73363817","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.2015.7280542
Haobin Dou, Xihong Wu
Convolutional Neural Networks (CNNs) have become forceful models in feature learning and image classification. They achieve translation invariance by spatial convolution and pooling mechanisms, while their ability in scale invariance is limited. To tackle the problem of scale variation in image classification, this work proposed a multi-scale CNN model with depth-decreasing multi-column structure. Input images were decomposed into multiple scales and at each scale image, a CNN column was instantiated with its depth decreasing from fine to coarse scale for model simplification. Scale-invariant features were learned by weights shared across all scales and pooled among adjacent scales. Particularly, a coarse-to-fine pre-training method imitating the human's development of spatial frequency perception was proposed to train this multi-scale CNN, which accelerated the training process and reduced the classification error. In addition, model averaging technique was used to combine models obtained during pre-training and further improve the performance. With these methods, our model achieved classification errors of 15.38% on CIFAR-10 dataset and 41.29% on CIFAR-100 dataset, i.e. 1.05% and 2.97% reduction compared with single-scale CNN model.
{"title":"Coarse-to-fine trained multi-scale Convolutional Neural Networks for image classification","authors":"Haobin Dou, Xihong Wu","doi":"10.1109/IJCNN.2015.7280542","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280542","url":null,"abstract":"Convolutional Neural Networks (CNNs) have become forceful models in feature learning and image classification. They achieve translation invariance by spatial convolution and pooling mechanisms, while their ability in scale invariance is limited. To tackle the problem of scale variation in image classification, this work proposed a multi-scale CNN model with depth-decreasing multi-column structure. Input images were decomposed into multiple scales and at each scale image, a CNN column was instantiated with its depth decreasing from fine to coarse scale for model simplification. Scale-invariant features were learned by weights shared across all scales and pooled among adjacent scales. Particularly, a coarse-to-fine pre-training method imitating the human's development of spatial frequency perception was proposed to train this multi-scale CNN, which accelerated the training process and reduced the classification error. In addition, model averaging technique was used to combine models obtained during pre-training and further improve the performance. With these methods, our model achieved classification errors of 15.38% on CIFAR-10 dataset and 41.29% on CIFAR-100 dataset, i.e. 1.05% and 2.97% reduction compared with single-scale CNN model.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"7 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":"77134962","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.2015.7280393
Abbas Shahzadeh, A. Khosravi, S. Nahavandi
Utility companies provide electricity to a large number of consumers. These companies need to have an accurate forecast of the next day electricity demand. Any forecast errors will result in either reliability issues or increased costs for the company. Because of the widespread roll-out of smart meters, a large amount of high resolution consumption data is now accessible which was not available in the past. This new data can be used to improve the load forecast and as a result increase the reliability and decrease the expenses of electricity providers. In this paper, a number of methods for improving load forecast using smart meter data are discussed. In these methods, consumers are first divided into a number of clusters. Then a neural network is trained for each cluster and forecasts of these networks are added together in order to form the prediction for the aggregated load. In this paper, it is demonstrated that clustering increases the forecast accuracy significantly. Criteria used for grouping consumers play an important role in this process. In this work, three different feature selection methods for clustering consumers are explained and the effect of feature extraction methods on forecast error is investigated.
{"title":"Improving load forecast accuracy by clustering consumers using smart meter data","authors":"Abbas Shahzadeh, A. Khosravi, S. Nahavandi","doi":"10.1109/IJCNN.2015.7280393","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280393","url":null,"abstract":"Utility companies provide electricity to a large number of consumers. These companies need to have an accurate forecast of the next day electricity demand. Any forecast errors will result in either reliability issues or increased costs for the company. Because of the widespread roll-out of smart meters, a large amount of high resolution consumption data is now accessible which was not available in the past. This new data can be used to improve the load forecast and as a result increase the reliability and decrease the expenses of electricity providers. In this paper, a number of methods for improving load forecast using smart meter data are discussed. In these methods, consumers are first divided into a number of clusters. Then a neural network is trained for each cluster and forecasts of these networks are added together in order to form the prediction for the aggregated load. In this paper, it is demonstrated that clustering increases the forecast accuracy significantly. Criteria used for grouping consumers play an important role in this process. In this work, three different feature selection methods for clustering consumers are explained and the effect of feature extraction methods on forecast error is investigated.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"9 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":"82143701","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.2015.7280494
Yongjin Kwon, K. Kang, C. Bae
It has been more important to measure daily physical activity for several purposes. There have been a number of methods of measuring physical activity, such as self-reporting, attaching wearable sensors, etc. Since a smartphone has become widespread rapidly, physical activity can be easily measured by accelerometers in the smartphone. Although there were a number of studies for activity recognition exploiting smartphone acceleration data, there was little discussion with the influence of each axis of accelerometers for activity recognition. In this paper, we investigate how each axis of smartphone acceleration data is affected on the performance of human activity recognition using a neural network based classifier. Assuming that the smartphone is kept in a pants pocket, the acceleration data of a subject are collected during standing, walking, and running for ten minutes. A multilayer perceptron was used as an activity classifier to recognize the three activities. Using averages as features, the classifier with the x-axis features provides the best accuracies. Using standard deviations as features, however, the accuracies are better than those using averages.
{"title":"Analysis and evaluation of smartphone-based human activity recognition using a neural network approach","authors":"Yongjin Kwon, K. Kang, C. Bae","doi":"10.1109/IJCNN.2015.7280494","DOIUrl":"https://doi.org/10.1109/IJCNN.2015.7280494","url":null,"abstract":"It has been more important to measure daily physical activity for several purposes. There have been a number of methods of measuring physical activity, such as self-reporting, attaching wearable sensors, etc. Since a smartphone has become widespread rapidly, physical activity can be easily measured by accelerometers in the smartphone. Although there were a number of studies for activity recognition exploiting smartphone acceleration data, there was little discussion with the influence of each axis of accelerometers for activity recognition. In this paper, we investigate how each axis of smartphone acceleration data is affected on the performance of human activity recognition using a neural network based classifier. Assuming that the smartphone is kept in a pants pocket, the acceleration data of a subject are collected during standing, walking, and running for ten minutes. A multilayer perceptron was used as an activity classifier to recognize the three activities. Using averages as features, the classifier with the x-axis features provides the best accuracies. Using standard deviations as features, however, the accuracies are better than those using averages.","PeriodicalId":6539,"journal":{"name":"2015 International Joint Conference on Neural Networks (IJCNN)","volume":"56 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":"78868227","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：