KonicaMinolta Medical and Graphic Inc.

Hachiōji, Japan

KonicaMinolta Medical and Graphic Inc.

Hachiōji, Japan
Time filter
Source Type

Tanaka J.,Saitama University | Nagashima M.,Saitama University | Kido K.,KonicaMinolta Medical and Graphic Inc. | Hoshino Y.,KonicaMinolta Medical and Graphic Inc. | And 5 more authors.
Zeitschrift fur Medizinische Physik | Year: 2013

We developed an X-ray phase imaging system based on Talbot-Lau interferometry and studied its feasibility for clinical diagnoses of joint diseases. The system consists of three X-ray gratings, a conventional X-ray tube, an object holder, an X-ray image sensor, and a computer for image processing. The joints of human cadavers and healthy volunteers were imaged, and the results indicated sufficient sensitivity to cartilage, suggesting medical significance. © 2012.

Momose A.,Tohoku University | Yashiro W.,Tohoku University | Kido K.,KonicaMinolta Medical and Graphic Inc. | Kiyohara J.,KonicaMinolta Medical and Graphic Inc. | And 9 more authors.
Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences | Year: 2014

With the aim of clinical applications of X-ray phase imaging based on TalbotLau-type grating interferometry to joint diseases and breast cancer, machines employing a conventional X-ray generator have been developed and installed in hospitals. The machine operation especially for diagnosing rheumatoid arthritis is described, which relies on the fact that cartilage in finger joints can be depicted with a dose of several milligray. The palm of a volunteer observed with 19 s exposure (total scan time: 32 s) is reported with a depicted cartilage feature in joints. This machine is now dedicated for clinical research with patients. © 2014 The Authors.

Kiyohara J.,KonicaMinolta Medical and Graphic Inc. | Makifuchi C.,KonicaMinolta Medical and Graphic Inc. | Kido K.,KonicaMinolta Medical and Graphic Inc. | Nagatsuka S.,KonicaMinolta Medical and Graphic Inc. | And 6 more authors.
AIP Conference Proceedings | Year: 2012

The prototype of the Talbot-Lau interferometer for clinical use was designed and applied to the preclinical examination. The human cadaveric hand and the mastectomy specimen were imaged. As a result, the images obtained by Talbot-Lau interferometry sensitively depicted the cartilages or the intraductal carcinoma. This result indicated that the Talbot-Lau interferometry would be a promising technology of the image diagnosis. © 2012 American Institute of Physics.

Yamazaki K.,Tokyo Institute of Technology | Kaji D.,Konicaminolta Medical and Graphic Inc.
Neural Networks | Year: 2013

Hierarchical learning models are ubiquitously employed in information science and data engineering. The structure makes the posterior distribution complicated in the Bayes method. Then, the prediction including construction of the posterior is not tractable though advantages of the method are empirically well known. The variational Bayes method is widely used as an approximation method for application; it has the tractable posterior on the basis of the variational free energy function. The asymptotic behavior has been studied in many hierarchical models and a phase transition is observed. The exact form of the asymptotic variational Bayes energy is derived in Bernoulli mixture models and the phase diagram shows that there are three types of parameter learning. However, the approximation accuracy or interpretation of the transition point has not been clarified yet. The present paper precisely analyzes the Bayes free energy function of the Bernoulli mixtures. Comparing free energy functions in these two Bayes methods, we can determine the approximation accuracy and elucidate behavior of the parameter learning. Our results claim that the Bayes free energy has the same learning types while the transition points are different. © 2013 Elsevier Ltd.

Kaji D.,Tokyo Institute of Technology | Kaji D.,Konicaminolta Medical and Graphic INC. | Watanabe S.,Tokyo Institute of Technology
Neurocomputing | Year: 2011

Variational Bayes learning or mean field approximation is widely used in statistical models which are made of mixtures of exponential distributions, for example, normal mixtures, binomial mixtures, and hidden Markov models. To derive variational Bayes learning algorithm, we need to determine the hyperparameters in the a priori distribution; however, the design method of hyperparameters has not yet been established. In the present paper, we propose two different design methods of hyperparameters which are applied to the different purposes. In the former method, the hyperparameter is determined for minimization of the generalization error. In the latter method, it is chosen so that candidates of hidden structure in training data are extracted. It is experimentally shown that the optimal hyperparameters for two purposes are different from each other. © 2011 Elsevier B.V.

Loading KonicaMinolta Medical and Graphic Inc. collaborators
Loading KonicaMinolta Medical and Graphic Inc. collaborators