Time filter

Source Type

Sasano Y.,Osaka Prefecture University | Sasano Y.,Omron Software Co Ltd | Kise K.,Osaka Prefecture University
ICIC Express Letters | Year: 2012

Most of the conventional methods for object recognition are based on appearance models. In this paper, we propose a novel method for recognizing objects by observing human actions instead of object appearances, such as shapes and colors. Our method is based on a bag-of-features approach, in which n-grams of human poses are employed as features. First, HOG features of human poses while taking actions on a object are extracted from video images and then clustered to number of symbols. Then, n-grams are generated from the sequence of symbols and registered for corresponding object category. In order to recognize unknown object, actions taken on the object are converted into n-grams in the same way and compared with ones representing learned object. We performed experiments to recognize objects in an office environment and confirmed the effectiveness of our method.


Iwata K.,Aichi University | Nakashima T.,Sugiyama Jogakuen University | Anan Y.,Omron Software Co. | Ishii N.,Aichi Institute of Technology
Studies in Computational Intelligence | Year: 2012

In this paper, we cluster and analyze data from the past embedded software development projects using self-organizing maps (SOMs)[9] that are a type of artificial neural networks that rely on unsupervised learning. The purpose of the clustering and analysis is to improve the accuracy of predicting the number of errors. A SOMproduces a low-dimensional, discretized representation of the input space of training samples; these representations are called maps. SOMs are useful for visualizing low-dimensional views of high-dimensional data, a multidimensional scaling technique. The advantages of SOMs for statistical applications are as follows: (1) data visualization, (2) information processing on association and recollection, (3) summarizing large-scale data, and (4) creating nonlinear models. To verify our approach, we perform an evaluation experiment that compares SOM classification to product type classification using Welch's t-test for Akaike's Information Criterion (AIC). The results indicate that the SOM classification method is more contributive than product type classification in creating estimation models, because the mean AIC of SOM classification is statistically significantly lower.


Ohmura R.,Toyohashi University of Technology | Uchida R.,OMRON SOFTWARE Co.
UbiComp 2014 - Adjunct Proceedings of the 2014 ACM International Joint Conference on Pervasive and Ubiquitous Computing | Year: 2014

Disrupting the transmission of sensor data due to sensor failure or connection loss significantly decrease accuracy in existing activity recognition techniques. We introduce an approach towards managing missing sensor data which operates at each step of the standard activity recognition, beginning with raw sensor data, feature calculation, classification, and result, as well as their combination methods. Our evaluation showed that the F1-score increased from 0.61 in the case of sensor data loss to 0.68 with the combination of all methods. Moreover, by selecting the combination of methods according to the failed sensor position, the F1-score increased to 0.69. Copyright 2014 ACM.


Iwata K.,Aichi University | Nakashima T.,Sugiyama Jogakuen University | Anan Y.,Omron Software Co. | Ishii N.,Aichi Institute of Technology
Proceedings - International Conference on Tools with Artificial Intelligence, ICTAI | Year: 2011

In this paper, we create effort prediction models using self-organizing maps (SOMs) [1] for embedded software development projects. SOMs are a type of artificial neural networks that rely on unsupervised learning. They produce a low-dimensional, discretized representation of the input space of training samples; these representations are called maps. SOMs are useful for visualizing low-dimensional views of high-dimensional data a multidimensional scaling technique. The advantages of using SOMs for statistical applications are as follows: (1) enabling reasonable inferences to be made from incomplete information via association and recollection, (2) visualizing data, (3) summarizing large-scale data, and (4) creating nonlinear models. We focus on the first advantage to create effort prediction models. To verify our approach, we perform an evaluation experiment that compares SOM models to feedforward artificial neural network (FANN) models using Welch's t test. The results of the comparison indicate that SOM models are more accurate than FANN models for the mean of absolute errors when predicting the amount of effort, because mean errors of the SOM are statistically significantly lower. © 2011 IEEE.


Iwata K.,Aichi University | Nakashima T.,Sugiyama Jogakuen University | Anan Y.,Omron Software Co. | Ishii N.,Aichi Institute of Technology
Studies in Computational Intelligence | Year: 2013

In this study, we establish error prediction models at various stages of embedded software development using hybrid methods of self-organizing maps (SOMs) and multiple regression analyses (MRAs). SOMs are a type of artificial neural networks that relies on unsupervised learning. A SOM produces a low-dimensional, discretized representation of the input space of training samples; these representations are called maps. SOMs are useful for visualizing low-dimensional views of high-dimensional data as a multidimensional scaling technique. The advantages of SOMs for statistical applications are as follows: (1) enabling reasonable inferences to be made from incomplete information via association and recollection, (2) visualizing data, (3) summarizing large-scale data, and (4) creating nonlinear models. We focus on the first advantage to create error prediction models at various stages of embedded software development. In some cases, a model using only SOMs yields lower error prediction accuracy than a model using only MRAs. However, the opposite is true. Therefore, in order to improve prediction accuracy, we combine both models. To verify our approach, we perform an evaluation experiment that compares hybrid models to MRA models using Welch's t test. The results of the comparison indicate that the hybrid models are more accurate than the MRA models for the mean of relative errors, because the mean errors of the hybrid models are statistically significantly lower. © 2013 Springer-Verlag Berlin Heidelberg.


Yamamoto K.,Kyoto Institute of Technology | Kuriyama T.,OMRON Software Co. | Shigemori H.,Kyoto Municipal Transportation | Kuramoto I.,Kyoto Institute of Technology | And 2 more authors.
Proceedings - International Computer Software and Applications Conference | Year: 2011

Most of traditional retrieval keys are the place (folder), name and content of desired files. If a user does not remember these keys, it is difficult for him or her to access desired files. In order to solve this problem, we propose a file retrieval system that is based on provenance. In this paper, provenance means the history of the file's moving and editing that a user has experienced. By recording how and from where a user gets the file, the system enables the user to query, for example, "which files did I get from my teacher yesterday?", "which files did I download from the web page of my college last month?", and "which files did I edit by MS Word last week?" We present a model of provenance and describe how to use provenance as the keys of retrieval and how the system is constructed. © 2011 IEEE.


Iwata K.,Aichi University | Nakashima T.,Sugiyama Jogakuen University | Anan Y.,Omron Software Co. | Ishii N.,Aichi Institute of Technology
Studies in Computational Intelligence | Year: 2010

In this paper we propose a method for reducing the margin of error in effort and error prediction models for embedded software development projects using artificial neural networks(ANNs). In addition, we perform an evaluation experiment that uses Welch's t-test to compare the accuracy of the proposed ANN method with that of our original ANN model. The results show that the proposed ANN model is more accurate than the original one in predicting the number of errors for new projects, since the means of the errors in the proposed ANN are statistically significantly lower. © 2010 Springer-Verlag Berlin Heidelberg.


Iwata K.,Aichi University | Nakashima T.,Sugiyama Jogakuen University | Anan Y.,Omron Software Co. | Ishii N.,Aichi Institute of Technology
IEEJ Transactions on Electronics, Information and Systems | Year: 2010

In this paper, we establish effort and error prediction models using an artificial neural networks (ANNs). We propose the normalizing method to reduce the margin of errors for ANN models. In addition, we perform an evaluation experiment to compare the accuracy of the ANN models with that of the regression analysis (RA) model and that of two ANN models using Steel-Dwass's multiple comparison test. The results show that each ANN model is more accurate than the RA model and the proposed method can reduce the errors for some cases, since the mean errors of the ANN models are statistically significantly lower. © 2010 The Institute of Electrical Engineers of Japan.


Anan Y.,Omron Software Co. | Nakashima T.,Nagoya University | Iwata K.,Aichi University | Yonemitsu H.,Omron Software Kyusyu Co. | And 2 more authors.
IEEJ Transactions on Electronics, Information and Systems | Year: 2010

In this paper, we propose an errors estimation model for embedded software development projects and implement a model visual tool for processing projects with upper and lower limits from estimated errors. The models are derived by multiple regression analysis, because this way is well known and easily used by developers. In addition, we perform evaluate to conform the models effectiveness. © 2010 The Institute of Electrical Engineers of Japan.

Loading Omron Software Co Ltd collaborators
Loading Omron Software Co Ltd collaborators