Institution
Mitsubishi
Company•Tokyo, Japan•
About: Mitsubishi is a company organization based out in Tokyo, Japan. It is known for research contribution in the topics: Layer (electronics) & Signal. The organization has 53115 authors who have published 54821 publications receiving 870150 citations. The organization is also known as: Mitsubishi Group of Companies & Mitsubishi Companies.
Papers published on a yearly basis
Papers
More filters
••
27 Jun 2016TL;DR: A novel deep network, which is referred to as Gaussian Mean Field (GMF) network, whose layers perform mean field inference over a Gaussian CRF, which outperforms various recent semantic segmentation approaches that combine CNNs with discrete CRF models.
Abstract: In contrast to the existing approaches that use discrete Conditional Random Field (CRF) models, we propose to use a Gaussian CRF model for the task of semantic segmentation. We propose a novel deep network, which we refer to as Gaussian Mean Field (GMF) network, whose layers perform mean field inference over a Gaussian CRF. The proposed GMF network has the desired property that each of its layers produces an output that is closer to the maximum a posteriori solution of the Gaussian CRF compared to its input. By combining the proposed GMF network with deep Convolutional Neural Networks (CNNs), we propose a new end-to-end trainable Gaussian conditional random field network. The proposed Gaussian CRF network is composed of three sub-networks: (i) a CNN-based unary network for generating unary potentials, (ii) a CNN-based pairwise network for generating pairwise potentials, and (iii) a GMF network for performing Gaussian CRF inference. When trained end-to-end in a discriminative fashion, and evaluated on the challenging PASCALVOC 2012 segmentation dataset, the proposed Gaussian CRF network outperforms various recent semantic segmentation approaches that combine CNNs with discrete CRF models.
142 citations
••
142 citations
••
TL;DR: Abnormal melatonin secretion may be involved in postoperative sleep disturbances, which triggered delirium in elderly patients, and in 5 patients with complications, melatonin levels were markedly increased.
Abstract: Background: Melatonin, a hormone produced in the pineal gland, is involved in circadian rhythms and the sleep-wake cycle. Postoperative delirium is encountered frequently in elderly patients after major surgery; whether changes in the pattern of melatonin secretion are associated is unclear. Methods: Plasma samples were obtained every 2 hours from 19 patients without delirium and 10 with delirium after major abdominal surgery. Postoperative delirium was determined using the Confusion Assessment Method in the Practice Guideline of the American Psychiatric Association. Results: All patients without delirium showed nearly identical preoperative and postoperative melatonin secretion for 24 hours, although peak values were significantly lower in patients more than 80 years old (7.2 ± 2.3 pg/mL) than in patients younger than 80 years (24.4 ± 4.1 pg/mL, P = 0.022). Patients with delirium showed two different abnormal postoperative patterns: in 5 patients without complications, melatonin levels were lower than preoperative values (11.0 ± 5.8 versus 6.5 ± 4.2 pg/mL, P = 0.079); and in 5 patients with complications, melatonin levels were markedly increased (21.1 ± 4.5 versus 58.8 ± 12.4 pg/mL, P = 0.043). Conclusions: Abnormal melatonin secretion may be involved in postoperative sleep disturbances, which triggered delirium in elderly patients.
142 citations
••
TL;DR: A method for range sensing that projects a single pattern of multiple slits of randomly distributed dots that propagated by exploiting the adjacency relationships to get an entire range image is described.
Abstract: A method for range sensing that projects a single pattern of multiple slits is described. Random dots are used to identify each slit. The random dots are given as randomly distributed cuts on each slit. Thus, each slit is divided into many small line segments. Segment matching between the image and the pattern is performed to obtain 3-D data. Using adjacency relations among slit segments, the false matches are reduced, and segment pairs whose adjacent segments correspond with each other are extracted and considered to be correct matches. Then, from the resultant matches, the correspondence is propagated by exploiting the adjacency relationships to get an entire range image. >
142 citations
••
TL;DR: In this article, a technique for minimizing the detent force using the finite element method is described, which is based on the phase difference between the two magnetic forces arising at both side edges of a stator core.
Abstract: This paper describes a technique for minimizing the detent force using the finite element method. The detent force of the whole stator core is the total of two magnetic forces arising at both side edges of a stator core. Computed results show the phase difference between the two magnetic forces. Therefore, we can cancel out the two forces by adjusting a stator length to minimize the detent force. The stator with the smooth formed edge shape is also contrived to reduce the detent force in the practical use. The detent force of this model is successfully minimized by proposed method.
142 citations
Authors
Showing all 53117 results
Name | H-index | Papers | Citations |
---|---|---|---|
Thomas S. Huang | 146 | 1299 | 101564 |
Kazunari Domen | 130 | 908 | 77964 |
Kozo Kaibuchi | 129 | 493 | 60461 |
Yoshimi Takai | 122 | 680 | 61478 |
William T. Freeman | 113 | 432 | 69007 |
Tadayuki Takahashi | 112 | 932 | 57501 |
Takashi Saito | 112 | 1041 | 52937 |
H. Vincent Poor | 109 | 2116 | 67723 |
Qi Tian | 96 | 1030 | 41010 |
Andreas F. Molisch | 96 | 777 | 47530 |
Takeshi Sakurai | 95 | 492 | 43221 |
Akira Kikuchi | 93 | 412 | 28893 |
Markus Gross | 91 | 588 | 32881 |
Eiichi Nakamura | 90 | 845 | 31632 |
Michael Wooldridge | 87 | 543 | 50675 |