Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)  

When a 3D scene is captured in several 2D images, a compact description (or parameters) of the 3D scene can be estimated from the images. Such an inference is formulated as the inverse of rendering computer graphics and is important for various applications, such as object recognition, inspection, and/or VR. In the present paper, we extend a photometric stereo method in such a way as to estimate the texture of the object in addition to previous estimation of parameters describing the objects and light sources. To do so, we need a realistic minimization method, combined with a method to obtain the Jacobian of the cost function with respect to the texture. We implemented this method and verified the validity of the framework using synthetic and real-world data.

DOI： 10.1007/978-3-031-06381-7_3

Scopus

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

Nonnegative Matrix Factorization (NMF) has attracted a great deal of attention as an effective technique for dimensionality reduction of large-scale nonnegative data. Given a nonnegative matrix, NMF aims to obtain two low-rank nonnegative factor matrices by solving a constrained optimization problem. The Hierarchical Alternating Least Squares (HALS) algorithm is a well-known and widely-used iterative method for solving such optimization problems. However, the original update rule used in the HALS algorithm is not well defined. In this paper, we propose a novel well-defined update rule of the HALS algorithm, and prove its global convergence in the sense of Zangwill. Unlike conventional globally-convergent update rules, the proposed one allows variables to take the value of zero and hence can obtain sparse factor matrices. We also present two stopping conditions that guarantee the finite termination of the HALS algorithm. The practical usefulness of the proposed update rule is shown through experiments using real-world datasets.

DOI： 10.1007/s10898-022-01167-7

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Authorship：Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Springer  

Formulated as an inverse problem of computer graphics, inverse rendering (or photometric stereo) can estimate parameters, such as the shapes and configurations of objects and light sources in the scene, from a set of images. Such results can be useful for applications including recognition, inspection, and/or VR. In the present paper, we extend previous studies in such a way as to incorporate more complex image formation models, specifically superquadrics with cast shadows, that are implemented on a standard graphics API and verify the framework on synthetic and real-world data.

DOI： 10.1007/978-3-030-81638-4_22

Web of Science

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

Principal component analysis is one of the most important methods of multivariate analysis, and has been applied in a wide range of fields such as statistical analysis, machine learning, pattern recognition, signal processing, and communication. Recently, using the idea of multi-agent networks, distributed algorithms for principal component analysis have been proposed for the case where the data matrix is partitioned either row-wise or column-wise. In this paper, considering the case where the data matrix is partitioned both row-wise and column-wise, we propose a new algorithm that allows a multi-agent network to perform principal component analysis in a distributed manner. We also verify its validity by numerical experiments. The proposed algorithm is based on the power method for principal component analysis and the average consensus algorithm.

DOI： 10.1109/PIC50277.2020.9350816

Scopus

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

The problem of finding a simple regular graph with the specified order and degree that minimizes the average shortest path length has a long history in graph theory. Recently this problem has attracted a great deal of attention in relation to the design of computer networks in data centers. In this paper, we propose a genetic algorithm for finding an approximate solution to this problem. Because the search space is the set of all simple regular graphs with the specified order and degree, conventional genetic algorithms cannot be directly applied. We propose in this paper new crossover and mutation operators that guarantee the simplicity and regularity of graphs. We also evaluate the effectiveness of the proposed method experimentally.

DOI： 10.1109/SSCI47803.2020.9308205

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ssci/ssci2020.html#HayashiMT20

Language：English   Publishing type：Research paper (international conference proceedings)  

Matrix factorization is a popular technique used in recommender systems based on collaborative filtering. Given a matrix that represents ratings of items by users, one can obtain latent feature vectors of the users and the items by applying one of the existing matrix factorization algorithms. In this paper, we focus our attention on matrices obtained from categorical ratings using one-hot encoding, and propose an element-wise alternating least squares algorithm to obtain latent feature vectors from such matrices. We next show that the proposed algorithm has the global convergence property in the sense of Zangwill. We also show through experiments using a benchmark dataset that the proposed algorithm is effective for prediction of unknown ratings.

DOI： 10.1007/978-3-030-63823-8_40

Scopus

researchmap

Authorship：Corresponding author   Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Springer  

Inverse rendering is a method for estimating object shapes and lighting conditions from a set of images. In the present paper, we propose an uncalibrated photometric stereo method that approximates the shapes of the target objects using quadric surfaces (ellipsoids or hyperboloids), and this shape representation allows us to use multiple cameras in order to reduce the GBR ambiguity. The proposed method uses the Levenberg-Marquardt method to minimize the differences between the input images and generated images calculated based on the estimated parameters and the image formation model. It is important to define each term in the cost function depending on the inside or outside of the contour of the estimated quadric surface and/or the target object. The proposed method has been verified on several sets of real images, namely, Vase, Grapes, and Snacks, and we found that the method required less than 100 iterations. On the other hand, there is room for improvement regarding the accuracy of the estimation by taking into account the shadows of the target objects and using a more advanced reflection model.

DOI： 10.1007/978-981-15-4818-5_24

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/fcv/fcv2020.html#NasuMST20

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

In some applications of multiagent networks, it is desired that each agent can evaluate how well the network is connected. In this paper, we propose a novel discrete-time algorithm for each agent to compute the algebraic connectivity of the network in a pseudo-decentralized manner. The proposed algorithm requires less computational cost than the conventional algorithm. We also analyze the dynamical behavior of the proposed algorithm and prove under some assumptions on the parameter values and the initial state values of the agents that all agents can compute the algebraic connectivity.

DOI： 10.1109/SSCI44817.2019.9002875

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ssci/ssci2019.html#ShimadaMT19

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

This paper proposes a distributed algorithm for multiple agents to perform the Nonnegative Matrix Factorization (NMF) based on the Euclidean distance. The matrix to be factorized is partitioned into multiple blocks, and each block is assigned to one of the agents forming a two-dimensional grid network. Each agent handles a small number of entries of the factor matrices corresponding to the assigned block, and updates their values by using information coming from the neighbors. It is shown that the proposed algorithm simulates the hierarchical alternating least squares method, which is well known as a fast algorithm for NMF based on the Euclidean distance, by making use of a finite-time distributed consensus algorithm.

DOI： 10.1109/SSCI44817.2019.9003158

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ssci/ssci2019.html#DomenMT19

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE  

DOI： 10.1109/ICSAI48974.2019.9010306

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icsai/icsai2019.html#SanoMT19

Authorship：Lead author   Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Authorship：Lead author   Language：English   Publishing type：Research paper (international conference proceedings)  

researchmap

Authorship：Lead author   Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Springer  

We address a photometric stereo problem that has unknown lighting conditions. To estimate the shape, reflection properties, and lighting conditions, we employ a nonlinear minimization that searches for parameters that can synthesize images that best fit the input images. A similar approach has been reported previously, but it suffers from slow convergence due to specular reflection parameters. In this paper, we introduce specular-free residual minimization that avoids the negative effects of specular reflection components by projecting the residual onto the complementary space of the light color. The minimization process simultaneously searches for the optimal light color and other parameters. We demonstrate the effectiveness of the proposed method using several real and synthetic image sets.

DOI： 10.1007/978-3-642-25367-6_16

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/psivt/psivt2011-1.html#MigitaSS11

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

Tracking 3d pose of a known object is one of the most important problems in computer vision. This paper proposes an appearance-based approach to this problem by combining the sparse template matching and the particle filter. Although the combination of them has already been discussed for 2d tracker, it has not been applied for efficient 3d tracking. This paper discusses an appearance-based tracker when a surface model of the target 3d object and the initial pose are given. The fundamental framework of the particle filter is provided at first for implementing a pose tracker based on sparse 3d template matching. Then, the coarse-to-fine approach is introduced for efficient implementation. Although the fundamental particle filter often requires a lot of particles for sufficient tracking, the number of particles can be effectively reduced by the coarse-to-fine strategy. Experimental results for both the simulation data and real images show how the proposed method works in frame rate on 3GHz Core 2 Quad(single thread) without using GPU or other special hardware.

DOI： 10.1109/ICIG.2009.138

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icig/icig2009.html#OkaKMS09

Authorship：Lead author   Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：Springer  

Given a set of images captured with a fixed camera while a point light source moves around an object, we can estimate the shape, reflectance property and texture of the object, as well as the positions of the light source. Our formulation is a large-scale nonlinear optimization that allows us to adjust the parameters so that the images synthesized from all of the parameters optimally fit the input images. This type of optimization, which is a variation of the bundle adjustment for structure and motion reconstruction, is often employed to refine a carefully constructed initial estimation. However, the initialization task often requires a great deal of labor, several special devices, or both. In the present paper, we describe (i) an easy method of initialization that does not require any special devices or a precise calibration and (ii) an efficient algorithm for the optimization. The efficiency of the optimization method enables us to use a simple initialization. For a set of synthesized images, the proposed method decreases the residual to zero. In addition, we show that various real objects, including toy models and human faces, can be successfully recovered.

DOI： 10.1007/978-3-540-88690-7_31

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/eccv/eccv2008-3.html#MigitaOS08

Authorship：Lead author   Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

The best method for estimating the fundamental matrix and/or the epipole over a given set of point correspondences between two images is a nonlinear minimization, which searches a rank-2 fundamental matrix that minimizes the geometric error cost function. When convenience is preferred to accuracy, we often use a linear approximation method, which searches a rank-3 matrix that minimizes the algebraic error. Although it has been reported that the algebraic error causes very poor results, it is currently thought that the relatively inaccurate results of a linear estimation method are a consequence of neglecting the rank-2 constraint, and not a result of exploiting the algebraic error. However, the reason has not been analyzed fully. In the present paper, we analyze the effects of the cost function selection and the rank-2 constraint based on covariance matrix analyses and show theoretically and experimentally that it is more important to enforce the rank-2 constraint than to minimize the geometric cost function.

DOI： 10.1109/CVPR.2007.383116

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/cvpr/cvpr2007.html#MigitaS07

Language：English   Publishing type：Research paper (international conference proceedings)   Publisher：IEEE Computer Society  

This paper discusses a face recognition system for a dialogue interface robot that really works in ubiquitous environments and reports an experimental result of real-life test in a ubiquitous environment. While a central module of the face recognition system is composed of the decomposed eigenface method, the system also includes a special face detection module and the face registration module. Since face recognition should work on images. captured by a camera equipped on the interface robot, all the methods are tuned for the interface robot. The face detection and recognition modules accomplish robust face detection and recognition when one of the registered users is talking to the robot. Some interesting results are reported with careful analysis of a sufficient real-life experiment.

DOI： 10.1109/ICPR.2006.157

Web of Science

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icpr/icpr2006-3.html#SakaueKMS06

Authorship：Lead author   Language：English   Publishing type：Research paper (international conference proceedings)  

CiNii Article

researchmap

Language：Japanese  

researchmap

Publishing type：Research paper (international conference proceedings)  

3D recovery of object shape and camera motion from 2D image sequence is formulated as a non-linear optimization problem. Such a non-linear approach ensures the most precise solution, but it involves computational difficulties, i.e. complexity and instability. This paper describes a practical procedure to solve the non-linear optimization problem for object shape and camera motion recovery. First, we discuss the imaging model for large objects such as buildings; one shot and pseudo wide angle models, where the camera motion is modeled as a uniform motion along circular path with gazing at the center. Then, the generalized initial values for circular motion is proposed and "mean appearance ratio (MAR)" of feature points is defined as a measure of stability. Experimental results with synthetic images show that 3D recovery was stably performed from image sequences with MAR over 15%, quasi-stable with over 8.5% for one shot model, and stable with over 10.3%, quasi-stable with over 4. 4% for pseudo wide angle model. 3D shape and motion using real images were recovered successfully from MAR 17% for one shot model and 6% for pseudo wide angle one. These results have demonstrated the robustness of the generalized initial values and effectiveness of MAR for large object shape from uncalibrated camera motion problem.

Scopus

researchmap

Authorship：Lead author   Language：Japanese  

researchmap

Publishing type：Research paper (international conference proceedings)  

This paper describes a method for recovering entire 3D shape of a building from an image sequence. This is a kind of “Shape and Motion recovery” problem in the computer vision field, whereas the conventional methods do not work well with the images taken around a large object in the near distance. Since such image sequence is a set of partial observations, the problem of 3D recovery becomes unstable. We first discuss the property of the local minima of the nonlinear optimization function, and then describe an incremental procedure to find the global minimum by avoiding the pseudo solutions. Experiments using the real images have shown that the proposed method successfully recovered the 3D shapes from eleven sets of image sequences.

Scopus

researchmap

Publishing type：Research paper (international conference proceedings)  

researchmap

Other Link： https://dblp.uni-trier.de/rec/conf/mva/2000

Authorship：Lead author   Language：Japanese  

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

Simultaneous recovery of shape and motion from an image sequence is directly formulated as minimization of reprojection errors via the nonlinear least squares. Although several linear foromulations which ensure unique solution are also known, the nonlinear approach can be applied to much broader configurations, and gives more accurate results. We investigate fast optimization algorithm and several methods to avoid local minima for nonlinear reconstruction problem. We test our methods on recovering entire 3D shape from a building image sequence with several hundred images and several hundred feature points.

CiNii Article

CiNii Books

researchmap

Other Link： http://id.nii.ac.jp/1001/00052498/

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

We propose a procedure for recovering entire 3D shape from a building image sequence with non-linear optimization method. In this case, the initial values of the unknown variables in the optimization process is crucial. Thus the direct application of the non-linear optimization process has not been fully used in shape recovery. In this paper, first, we investigate an initialization method, which is very simple and relatively generic to the building shape recovery problem. Then, we propose a method for restoring the optimization process from a local minimum. Experimental results using real and synthetic image have shown that our algorithm stably recover 3D shapes.

CiNii Article

CiNii Books

researchmap

Other Link： http://id.nii.ac.jp/1001/00052536/

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

Other Link： https://dblp.uni-trier.de/rec/conf/mva/2002

Language：English   Publishing type：Research paper, summary (international conference)  

researchmap

Other Link： https://dblp.uni-trier.de/rec/conf/mva/2000

J-GLOBAL

researchmap

Language：Japanese   Publisher：Information Processing Society of Japan (IPSJ)  

The task of computer vision is to recover accurate 3-dimensional shape of objects from 2-dimensional images. In such a vision task, the observer is assumed to be outside of the scene, and this means that the scene configuration are never destroyed by the observre. In this paper, we propose a new paradigm "active recognition", in which the object information is obtained by cooperation of visual and physical systems which can reform the scene configuration. Experiments on the estimation of the center of gravity was performed with a robot arm and a camera that were used to push and track the object. The stable results have shown the effectiveness of the integration of visual and physical information.

CiNii Article

CiNii Books

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publisher：The Institute of Electronics, Information and Communication Engineers  

CiNii Article

CiNii Books

researchmap

Event date： 2021.10

Language：Japanese  

researchmap

Event date： 2021.10

Language：Japanese  

researchmap

Event date： 2021

Language：English   Presentation type：Oral presentation (general)  

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/fcv/fcv2021.html#NasuMT21

Event date： 2021

Language：Japanese  

researchmap

Event date： 2021

Language：Japanese  

researchmap

Event date： 2021

Language：Japanese  

researchmap

Event date： 2021

Language：Japanese  

researchmap

Event date： 2020.12

Language：English  

researchmap

Event date： 2020

Language：Japanese  

researchmap

Event date： 2020

Language：English  

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ssci/ssci2020.html#HayashiMT20

Event date： 2020

Language：English  

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/iconip/iconip2020-5.html#WuMT20

Event date： 2020

Language：English  

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/fcv/fcv2020.html#NasuMST20

Event date： 2020

Language：Japanese  

researchmap

Event date： 2020

Language：Japanese  

researchmap

Event date： 2020

Language：Japanese  

researchmap

Event date： 2020

Language：Japanese  

researchmap

Event date： 2020

Language：Japanese  

researchmap

Event date： 2020

Language：Japanese  

researchmap

Event date： 2020

Language：Japanese  

researchmap

Event date： 2019.12

Language：English  

researchmap

Event date： 2019.12

Language：English  

researchmap

Event date： 2019.2

Language：English  

researchmap

Event date： 2019

Language：Japanese  

researchmap

Event date： 2019

Language：English  

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ssci/ssci2019.html#DomenMT19

Event date： 2019

Language：English  

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/ssci/ssci2019.html#ShimadaMT19

Event date： 2019

Language：English  

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icsai/icsai2019.html#SanoMT19

Event date： 2019

Language：Japanese  

researchmap

Event date： 2019

Language：Japanese  

researchmap

Event date： 2019

Language：Japanese  

researchmap

Event date： 2019

Language：Japanese  

researchmap

Event date： 2018

researchmap

Event date： 2018

researchmap

Event date： 2017.8.7 - 2017.8.10

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：広島国際会議場  

researchmap

Event date： 2016.3.24

Language：Japanese  

researchmap

Event date： 2016.2.17 - 2016.2.19

Language：English   Presentation type：Oral presentation (general)  

Venue：Takayama  

researchmap

Event date： 2016

researchmap

Event date： 2015

researchmap

Event date： 2015

researchmap

Event date： 2013.1.23

Language：Japanese  

This paper discusses on face tracking and recognition when depth-image sensor is available. For the purpose, our image based approach is extended for covering depth sensor. Experimental results on real depth-image sequences show the effectiveness of the proposed method.

researchmap

Event date： 2013

researchmap

Event date： 2012.1.12

Language：Japanese  

researchmap

Event date： 2012

researchmap

Event date： 2011

researchmap

Event date： 2011

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/psivt/psivt2011-1.html#MigitaSS11

Event date： 2010

researchmap

Event date： 2010

researchmap

Event date： 2009

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icig/icig2009.html#OkaKMS09

Event date： 2008.3.10

Language：Japanese  

Given a set of images, where a point light source moves around an object, we can formulate recovery of the reflectance property, texture and shape of the object, as well as the positions of the light source. We herein describe an efficient algorithm to address this large-scale nonlinear optimization problem. Various objects, including toy models and human faces, are successfully recovered without any special devices nor a precise calibration.

researchmap

Event date： 2008.1.17

Language：Japanese  

This report gives an overview of 2007 IEEE Computer Society Conference on Compute Vision and Pattern Recognition(CVPR2007), held in Minneapolis, Minnesota, USA, from June 19th to 21st, 2007.

researchmap

Event date： 2008

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/eccv/eccv2008-3.html#MigitaOS08

Event date： 2008

researchmap

Event date： 2008

researchmap

Event date： 2007

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/cvpr/cvpr2007.html#MigitaS07

Event date： 2006.3.17

Language：Japanese  

The fundamental matrix can be estimated from 7 or more point correspondences in two uncalibrated views, and essentially it is an epipole search in 2-dimensional space (hemisphere). However, since the corresponding cost function is a non-linear one, unique solution is not guaranteed. While the well-known 8-point algorithm gives an approximate solution, neglecting the rank constraint on the fundamental matrix, we strictly take the constraint into account and obtain a better solution. Specifically, we reduce the problem into a high-order polynomial equation in one variable, which is satisfied by the ratio (e.g. x/z) of all the optimal or locally optimal epipole coordinates (x, y, z). We also show a lower-order equation by introducing an approximation, which still outperforms the 8-point algorithm.

researchmap

Event date： 2006

researchmap

Event date： 2006

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/psivt/psivt2006.html#MigitaS06

Event date： 2006

researchmap

Other Link： https://dblp.uni-trier.de/db/conf/icpr/icpr2006-3.html#SakaueKMS06

Event date： 2005.10.27

Language：Japanese  

The paper presents a face recognition system for a dialog interface robot in ubiquitous environments. In the system, a decomposed eigenface method is used for the face recognition. The method realizes robust recognition under various lighting conditions. We report experimental results in Ubiquitous Home using the system.

researchmap

Event date： 2005.1

Venue：Gwangju, Korea  

researchmap

Event date： 2005

researchmap

Event date： 2005

researchmap

Event date： 2003.6.11 - 2003.6.13

Venue：Halifax, Canada  

researchmap

Event date： 2002.7.30 - 2002.8.1

Venue：名古屋  

researchmap

Event date： 2002.5.27 - 2002.5.29

Venue：Calgary, Canada  

researchmap

Event date： 2002.5.27 - 2002.5.29

Venue：Calgary, Canada  

researchmap

Event date： 2001.12.14 - 2001.12.15

Venue：広島  

researchmap

Event date： 2001.1.13 - 2001.1.14

Venue：広島  

researchmap

Event date： 2000.7.18 - 2000.7.21

Venue：長野  

researchmap

Event date： 2000.7.18 - 2000.7.21

Venue：長野  

researchmap

Event date： 2000.1.18 - 2000.1.19

Venue：広島  

researchmap

Language：Japanese   Presentation type：Oral presentation (general)  

researchmap

Language：Japanese   Presentation type：Oral presentation (general)  

researchmap

Language：Japanese   Presentation type：Oral presentation (general)  

researchmap

Language：Japanese   Presentation type：Oral presentation (general)  

researchmap

Language：English   Presentation type：Oral presentation (general)  

researchmap

Language：English   Presentation type：Oral presentation (general)  

researchmap

Presentation type：Oral presentation (general)  

researchmap