掲載種別：研究論文（学術雑誌）   出版者・発行元：MDPI AG  

In this paper, we propose a method to detect Braille blocks from an egocentric viewpoint, which is a key part of many walking support devices for visually impaired people. Our main contribution is to cast this task as a multi-objective optimization problem and exploits both the geometric and the appearance features for detection. Specifically, two objective functions were designed under an evolutionary optimization framework with a line pair modeled as an individual (i.e., solution). Both of the objectives follow the basic characteristics of the Braille blocks, which aim to clarify the boundaries and estimate the likelihood of the Braille block surface. Our proposed method was assessed by an originally collected and annotated dataset under real scenarios. Both quantitative and qualitative experimental results show that the proposed method can detect Braille blocks under various environments. We also provide a comprehensive comparison of the detection performance with respect to different multi-objective optimization algorithms.

DOI： 10.3390/s21082775

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：WILEY  

The recent literature indicates that face recognition and facial expression classification has achieved a high accuracy on benchmark datasets with a large number of face images in the wild. On the other hand, unlike the purpose of recognizing as many people as possible, real applications for families or companies usually aim to recognize a small group of people as accurate as possible. In case of the face is partially occluded, convolutional solutions always simply put images with occlusions into the training dataset and hope the convolution neural network learn a model robust to partial occlusion. These processes not only increase the burden of learning, but also affect the model to identify normal images without occlusions. In this paper, we investigate the research problem of facial image reconstruction to discard the influence of partial occlusion. Based on the phenomenon that human faces are roughly symmetrical, we propose to reconstruct the facial information of occluded areas with the intact half face. Specifically, occlusion on the left-half face is reconstructed with a linear combination of features on the right-half face and vice versa. The process is modeled by keeping row sparsity for the coefficient matrix with l(2,1)-norm regularization while minimizing the reconstruction error. An alternative iterative algorithm is proposed to solve the optimization problem. To validate the effectiveness of the reconstruction, we fine-tune a pretrained model, AlexNet, training on normal face images and test with various occluded images. Extensive experimental results show that our method has improved the classification performance effectively. (c) 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

DOI： 10.1002/tee.23335

Web of Science

researchmap

掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：SPIE  

DOI： 10.1117/12.2590778

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：John Wiley and Sons Inc  

In this study, population-based meta-heuristic algorithms—artificial bee colony, differential evolution, particle swarm optimization, and real-coded genetic algorithm—are applied to affine template matching for performance comparison. It is necessary to optimize six parameters for affine template matching. This is a combinatorial optimization problem, and the number of candidate solutions is very large. For such a problem, population-based meta-heuristic algorithms can efficiently search a global optimum. There is research that applies the algorithms to affine template matching. However, they select a specific algorithm without understanding the characteristics of affine template matching and comparing different algorithms. This means the selected algorithm may not be suitable for affine template matching. Hence, this research first analyzes the characteristics of affine template matching and compares the performance of the four algorithms. In addition, we propose a new method to measure population diversity for performance comparison. Finally, we confirmed that artificial bee colony achieves the best performance of the four methods. © 2020 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

DOI： 10.1002/tee.23274

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：WILEY  

Most face detection and recognition tasks are based on the training of intact facial images and corresponding labels. Both the three-dimensional (3D) structure and two-dimensional (2D) appearance from the frontal view of human faces are bilaterally symmetrical in general. However, sometimes, illumination on the left and right halves of faces is uneven. In such cases, the symmetrical characteristic of human faces can facilitate expressing distinct identity information. This is because even if one side of the facial image is corrupted by noise, the opposite side can still be used for feature extraction. This paper proposes an automatic selection of the better half of the face using only a half-face for identity recognition. Unlike the MegaFace challenge of recognizing millions of identities in the wild, this paper focuses on building recognition systems for a small number of people with fewer training images; the recognition system can, for example, build access control systems for research laboratory members or family members. This paper proposes an artificial face image construction method and a half-face training strategy for transfer learning of pretrained conventional neural network models. Extensive experimental results show that the proposed method improves the performance of state-of-the-art models by utilizing the symmetrical characteristics of human faces. (c) 2020 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

DOI： 10.1002/tee.23273

Web of Science

researchmap

掲載種別：研究論文（学術雑誌）   出版者・発行元：Springer Science and Business Media LLC  

Abstract

Evolutionary algorithms (EAs) and swarm algorithms (SAs) have shown their usefulness in solving combinatorial and NP-hard optimization problems in various research fields. However, in the field of computer vision, related surveys have not been updated during the last decade. In this study, inspired by the recent development of deep neural networks in computer vision, which embed large-scale optimization problems, we first describe a literature survey conducted to compensate for the lack of relevant research in this area. Specifically, applications related to the genetic algorithm and differential evolution from EAs, as well as particle swarm optimization and ant colony optimization from SAs and their variants, are mainly considered in this survey.

DOI： 10.1186/s41074-020-00065-9

researchmap

その他リンク： https://link.springer.com/article/10.1186/s41074-020-00065-9/fulltext.html

掲載種別：研究論文（学術雑誌）   出版者・発行元：Wiley  

DOI： 10.1002/tee.23134

researchmap

その他リンク： https://onlinelibrary.wiley.com/doi/full-xml/10.1002/tee.23134

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：John Wiley and Sons Inc.  

Automobile surface defects, such as scratches and dents, can largely affect the first impressions of consumers. These defects are likely to occur during the processes of manufacturing and painting. Most global automobile companies still rely on visual inspection to detect the defects, which results in instability and inefficiency in the inspection procedure. In this paper, a low-cost detection system is proposed. The hardware part of the system consists of only light emitting diode (LED) tubes and a single camera. The software part of the system first obtains defect candidates greedily by thresholding binarization result. Then, candidates in each frame are filtered using a differential image. Finally, as the main contribution, our algorithm further filters the candidates using tracking trajectories of defect candidates over multiple frames. Experimental results demonstrate that our method is able to detect defects while suppressing false-positive detections and is effective in real-world situations. © 2020 Institute of Electrical Engineers of Japan. Published by John Wiley &amp
Sons, Inc.

DOI： 10.1002/tee.23096

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：Institute of Electrical and Electronics Engineers Inc.  

Well-known corner or local extrema feature based detectors such as FAST and DoG have achieved noticeable successes. However, detecting keypoints in the presence of blur has remained to be an unresolved issue. As a matter of fact, various kinds of blur (e.g., motion blur, out-of-focus and space-variant) remarkably increase challenges for keypoint detection. As a result, those methods have limited performance. To settle this issue, we propose a blur-countering method for detecting valid keypoints for various types and degrees of blurred images. Specifically, we first present a distance metric for derivative distributions, which preserves the distinctiveness of patch pairs well under blur. We then model the asymmetry by utilizing the difference of squared eigenvalues based on the distance metric. To make it scale-robust, we also extend it to scale space. The proposed detector is efficient as the main computational cost is the square of derivatives at each pixel. Extensive visual and quantitative results show that our method outperforms current approaches under different types and degrees of blur. Without any parallelization, our implementation achieves real-time performance for low-resolution images (e.g., 320 × 240 pixel).

DOI： 10.1109/ACCESS.2020.3020561

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC  

Symmetry is an important visual characteristic in the image and can be used to identify real-world objects based on their geometrically balanced structures. Although image analysis of axisymmetry has been studied for years and many approaches have been developed to detect axisymmetry, image analysis of center-symmetry has been received little attention. Symmetrical center detection is a very important inspection for image analysis. The detected result is useful for image processing, such as object detection, image inpainting. It is difficult to detect the symmetrical center for the real-world image. Because the symmetrical objects usually appear some geometric transformation making them not perfectly center-symmetrical. This article proposes a novel weight voting approach to detect the global symmetrical center from a single image, noted as weight voting for center-symmetrical object detection (WVCS). Firstly, we sampled some feature points to vote for the global symmetrical center. Then, the orientation of Log-Gabor response and color information of HSV space are utilized as the feature descriptor to compute a similarity measure for two points. A proposed penalty term is employed to focus on the center-symmetrical objects in the foreground. The position with the maximal weight voted by all sampled feature points represents the symmetrical center. Based on the detected symmetrical object can be detected by fitting the symmetrical feature point pairs. The experimental results show that the WVCS outperforms other state-of-the-art algorithms while detecting the symmetrical center and objects from real-world images.

DOI： 10.1109/ACCESS.2020.3043528

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：BMVA Press  

We propose a novel multi-scale template matching method which is robust against both scaling and rotation in unconstrained environments. The key component behind is a similarity measure referred to as scalable diversity similarity (SDS). Specifically, SDS exploits bidirectional diversity of the nearest neighbor (NN) matches between two sets of points. To address the scale-robustness of the similarity measure, local appearance and rank information are jointly used for the NN search. Furthermore, by introducing penalty term on the scale change, and polar radius term into the similarity measure, SDS is shown to be a well-performing similarity measure against overall size and rotation changes, as well as non-rigid geometric deformations, background clutter, and occlusions. The properties of SDS are statistically justified, and experiments on both synthetic and real-world data show that SDS can significantly outperform state-of-the-art methods.

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：Institute of Electrical and Electronics Engineers Inc.  

Recently, deep convolutional neural networks (CNN) have become a new standard in many machine learning applications not only in image but also in audio processing. However, most of the studies only explore a single type of training data. In this paper, we present a study on classifying bird species by combining deep neural features of both visual and audio data using kernel-based fusion method. Specifically, we extract deep neural features based on the activation values of an inner layer of CNN. We combine these features by multiple kernel learning (MKL) to perform the final classification. In the experiment, we train and evaluate our method on a CUB-200-2011 standard data set combined with our originally collected audio data set with respect to 200 bird species (classes). The experimental results indicate that our CNN+MKL method which utilizes the combination of both categories of data outperforms single-modality methods, some simple kernel combination methods, and the conventional early fusion method.

DOI： 10.1109/CW.2019.00022

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：John Wiley and Sons Inc.  

Content-based image retrieval technique is an essential component under various application scenarios. In this paper, instead of visual similarity defined by colors, shapes, or textures, we aim to retrieve images with respect to the visual similarity defined by the human pose. In our framework, all the poses are derived from images, inspired by the recent development of three-dimension (3D) human pose reconstruction. Furthermore, to make the retrieval more robust against reconstruction error, we propose a recurrent bidirectional similarity measure called recurrent best-buddies similarity (RBBS). Specifically, we treat the similarity measure between two visual poses as a distance measure between two point vectors, with each point representing one of the reconstructed 3D human pose candidates. We then recur the similarity measure by the displacement of query. As a justification, we verify the validity of RBBS in a one dimension (1D) Gaussian situation. In experiments, we build an original dataset for the retrieval task. Both the qualitative and quantitative results show the usefulness of our framework
the quantitative results evaluated by mean average precision (MAP or mAP) especially demonstrate that RBBS is improved by 14.13% compared to the most competitive alternative methods. © 2019 Institute of Electrical Engineers of Japan. Published by John Wiley &amp
Sons, Inc.

DOI： 10.1002/tee.22902

Scopus

researchmap

掲載種別：研究論文（学術雑誌）   出版者・発行元：Japan Society for Precision Engineering  

DOI： 10.2493/jjspe.85.98

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：ASSOC COMPUTING MACHINERY  

We propose a novel genetic algorithm to solve the image deformation estimation problem by preserving the genetic diversity. As a classical problem, there is always a trade-off between the complexity of deformation models and the difficulty of parameters search in image deformation. 2D cubic B-spline surface is a highly free-form deformation model and is able to handle complex deformations such as fluid image distortions. However, it is challenging to estimate an apposite global solution.To tackle this problem, we develop a genetic operation named probabilistic bitwise operation (PBO) to replace the crossover and mutation operations, which can preserve the diversity during generation iteration and achieve better coverage ratio of the solution space. Furthermore, a selection strategy named annealing selection is proposed to control the convergence. Qualitative and quantitative results on synthetic data show the effectiveness of our method.

DOI： 10.1145/3319619.3321890

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：Institute of Electrical and Electronics Engineers Inc.  

In the recent decade, the development of 3D scanners brings the expansion of 3D models, which yields in the increase of demand for developing effective 3D point cloud retrieval methods using only unorganized point clouds instead of mesh data. In this paper, we propose a meshing-free framework for point cloud retrieval by exploiting a bidirectional similarity measurement on local features. Specifically, we first introduce an effective pipeline for keypoint selection by applying principal component analysis to pose normalization and thresholding local similarity of normals. Then, a point cloud based feature descriptor is employed to compute local feature descriptors directly from point clouds. Finally, we propose a bidirectional feature match strategy to handle the similarity measure. Experimental evaluation on a publicly available benchmark demonstrates the effectiveness of our framework and shows it can outperform other alternatives involving state-of-the-art techniques.

DOI： 10.1109/ACCESS.2019.2952157

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：Institute of Electronics, Information and Communication, Engineers, IEICE  

In recent decade, many state-of-the-art algorithms on image classification as well as audio classification have achieved noticeable successes with the development of deep convolutional neural network (CNN). However, most of the works only exploit single type of training data. In this paper, we present a study on classifying bird species by exploiting the combination of both visual (images) and audio (sounds) data using CNN, which has been sparsely treated so far. Specifically, we propose CNN-based multimodal learning models in three types of fusion strategies (early, middle, late) to settle the issues of combining training data cross domains. The advantage of our proposed method lies on the fact that we can utilize CNN not only to extract features from image and audio data (spectrogram) but also to combine the features across modalities. In the experiment, we train and evaluate the network structure on a comprehensive CUB-200-2011 standard data set combing our originally collected audio data set with respect to the data species. We observe that a model which utilizes the combination of both data outperforms models trained with only an either type of data. We also show that transfer learning can significantly increase the classification performance.

DOI： 10.1587/transinf.2018EDP7383

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：SPIE  

In mobile robots and autonomous vehicles, the information of depth and camera's orientation are necessary for the applications of positioning and navigation. In this paper, we propose an appearance-assumption-based method for the detection of horizontal line (HL) or vanishing line (VL) on the image observed with a single omnidirectional camera. HL can provide the important information about depth and camera's orientation for the single view estimation. The detection of HL is assumed as an optimization problem. The optimization of the line is adopted by the genetic algorithm (GA). We implement the real-time detection method of HL using GA. Therefore, the proposed method is evaluated on the multiple videos with the various environment. Experimental results demonstrate the efficiency of the proposed method.

DOI： 10.1117/12.2521591

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：Institute of Electrical and Electronics Engineers Inc.  

Although template matching has been studied for decades, difficult problems still exist, such as template matching with projective transformation. Projective template matching is challenging not only because of its own hardness, but also owing to the high cost in creating ground truths. Focusing on the problem in creating ground truths, in this paper, we propose a novel method to automatically create ground truth by augmented reality (AR) technology. Specifically, when creating dataset, around each target image we attach AR markers. Then we shoot videos containing target image with the markers from different angles and backgrounds. Finally we locate the target images by detecting the markers. We evaluate our method by deviation, intersection over union (IoU), and necessary correction rate, indicating that our method is available.

DOI： 10.1109/MECATRONICS.2018.8495744

Scopus

researchmap

掲載種別：研究論文（国際会議プロシーディングス）  

DOI： 10.12792/icisip2018.048

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）   出版者・発行元：Japan Society for Food Engineering  

The goal of this study is to construct a novel robot system for constant quantity filling cans (fixed can-filling task) in fish-processing plants. In this paper, we will demonstrate first a novel algorithm for the proper combination of a grilled fillet’s body and tail part as defined by the combined weight. As a result, the discrimination of experienced workers based only on combinations of visual information produced a correct answer rate of approximately 50%
the correct answer rate rose to about 90% when the sensation of weight obtained by grasping the grilled fillets could be used. On the other hand, a rate of approximately 95% correct answers was obtained by the proper combination algorithm based on the proposed three-dimensional feature that could quantify grilled fillets. Secondly, we will demonstrate the experimental verification of our novel robot system to select a pair of grilled fillet’s body and tail parts using the proper combination algorithm. This robot system consists of three linear-motion sliders in the X-Y-Z direction with a customized hand mechanism based on SWITL technology. As a result, our novel robot system for fixed can-filling task requires approximately 30 sec for picking up and transporting a pair of grilled fillets to the can. The system had 75% working accuracy despite skilled workers scoring 90%. Although the proposed robot system should improve the working accuracy and its speed of movement, these results suggest that our novel robot system for fixed can-filling task by installing the proper combination discrimination algorithm could replace the work of skilled workers in fixed can-filling.

DOI： 10.11301/jsfe.18517

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：WILEY  

In this paper, we propose a high-speed multiview face localization and tracking method with a minimum bounding box using template matching with genetic algorithm (GA). In this method, a head is treated as a cylinder, and the multiview face can be represented using the development of lateral surface (two-dimensional (2D) model). The face can be localized by a template, which is generated from the model and corresponds to the target face direction. Processing is very fast because the parameters for both template generation and affine transformation are simultaneously optimized by GA. In the experiment, challenging 60-video sequences are created in a situation where subjects drastically move their faces in a room, using a standard computer and web camera. Then, the proposed method is applied to the sequences, and the performance is investigated. As a result, the proposed method achieves fast and accurate multiview face localization and tracking. (C) 2017 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

DOI： 10.1002/tee.22460

Web of Science

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：PUBLIC LIBRARY SCIENCE  

The optomotor response of animals is commonly used to measure their visual performance, e.g., rats of different genetically altered strains or various drug tests. With the presentation of stimuli using computer screens or projectors, the common idea focuses on measuring the eye movement or head and/or body movement to characterize changes of the head gaze. However, traditional methods rely on either the invasive fixation of animals, or the judgment of a human observer who reports the stimulus-tracking movements. In this paper, we propose a novel head gaze determination system to automatically track the head movement of rats without artificial markers. The experiments were done to demonstrate the process of optimizing parameters in image processing. As a result, the head angle curve of the proposed method is consistent with that of ground-truth data annotated manually according to predefined rules. Hence, the proposed method provides a simple, convenient, and objective solution to automatically generate the head gaze orientations from massive amounts of recorded data for further visual performance analysis.

DOI： 10.1371/journal.pone.0176633

Web of Science

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（その他学術会議資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG  

We address the problem of measuring matching similarity in terms of template matching. A novel method called two-side agreement learning (TAL) is proposed which learns the implicit correlation between two sets of multi-dimensional data points. TAL learns from a matching exemplar to construct a symmetric tree-structured model. Two points from source set and target set agree to form a two-side agreement (TA) pair if each point falls into the same leaf cluster of the model. In the training stage, unsupervised weak hyper-planes of each node are learned at first. After then, tree selection based on a cost function yields final model. In the test stage, points are propagated down to leaf nodes and TA pairs are observed to quantify the similarity. Using TAL can reduce the ambiguity in defining similarity which is hard to be objectively defined and lead to more convergent results. Experiments show the effectiveness against the state-of-the-art methods qualitatively and quantitatively.

DOI： 10.1587/transinf.2016EDP7233

Web of Science

researchmap

記述言語：日本語   掲載種別：研究論文（その他学術会議資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG  

In this paper, we address the problem of non-parametric template matching which does not assume any specific deformation models. In real-world matching scenarios, deformation between a template and a matching result usually appears to be non-rigid and non-linear. We propose a novel approach called local rigidity constraints ( LRC). LRC is built based on an assumption that the local rigidity, which is referred to as structural persistence between image patches, can help the algorithm to achieve better performance. A spatial relation test is proposed to weight the rigidity between two image patches. When estimating visual similarity under an unconstrained environment, high-level similarity ( e. g. with complex geometry transformations) can then be estimated by investigating the number of LRC. In the searching step, exhaustive matching is possible because of the simplicity of the algorithm. Global maximum is given out as the final matching result. To evaluate our method, we carry out a comprehensive comparison on a publicly available benchmark and show that our method can outperform the state-of-the-art method.

DOI： 10.1587/transinf.2015EDP7492

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG  

In this paper, we address the problem of projective template matching which aims to estimate parameters of projective transformation. Although homography can be estimated by combining key-point-based local features and RANSAC, it can hardly be solved with feature-less images or high outlier rate images. Estimating the projective transformation remains a difficult problem due to high-dimensionality and strong non-convexity. Our approach is to quantize the parameters of projective transformation with binary finite field and search for an appropriate solution as the final result over the discrete sampling set. The benefit is that we can avoid searching among a huge amount of potential candidates. Furthermore, in order to approximate the global optimum more efficiently, we develop a level-wise adaptive sampling ( LAS) method under genetic algorithm framework. With LAS, the individuals are uniformly selected from each fitness level and the elite solution finally converges to the global optimum. In the experiment, we compare our method against the popular projective solution and systematically analyse our method. The result shows that our method can provide convincing performance and holds wider application scope.

DOI： 10.1587/transinf.2016EDP7038

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（その他学術会議資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（その他学術会議資料等）  

researchmap

記述言語：日本語   出版者・発行元：一般社団法人 日本機械学会  

<p>In order to contribute to fisheries industry in Iwate which was damaged by the Great East Japan Earthquake, we have developed prototypes of robots, that assist resource management, fish catch, seafood processing or circulation. By culture of fishes and shellfishes, it is physical and economical burden for fisherman to clean sewage such as remaining feed or excrement deposited at the bottom of the sea(sea culture) or at a tank floor(land culture). To reduce the burden and to improve breeding environment, we have developed a prototype of an underwater robot to clean a bottom of tank. In this paper, the concepts, system design and developed robot are described.</p>

DOI： 10.1299/jsmermd.2016.1A2-17a5

CiNii Article

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG  

In recent years, many variants of key point based image descriptors have been designed for the image matching, and they have achieved remarkable performances. However, to some images, local features appear to be inapplicable. Since theses images usually have many local changes around key points compared with a normal image, we define this special image category as the image with local changes (IL). An IL pair (ILP) refers to an image pair which contains a normal image and its IL. ILP usually loses local visual similarities between two images while still holding global visual similarity. When an IL is given as a query image, the purpose of this work is to match the corresponding ILP in a large scale image set. As a solution, we use a compressed HOG feature descriptor to extract global visual similarity. For the nearest neighbor search problem, we propose random projection indexed KD-tree forests (rKDFs) to match ILP efficiently instead of exhaustive linear search. rKDFs is built with large scale low-dimensional KD-trees. Each KD-tree is built in a random projection indexed subspace and contributes to the final result equally through a voting mechanism. We evaluated our method by a benchmark which contains 35,000 candidate images and 5,000 query images. The results show that our method is efficient for solving local-changes invariant image matching problems.

DOI： 10.1587/transinf.2015EDP7093

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

DOI： 10.1109/SOCPAR.2015.7492767

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：IEICE-INST ELECTRONICS INFORMATION COMMUNICATIONS ENG  

Tracking algorithms for arbitrary objects are widely researched in the field of computer vision. At the beginning, an initialized bounding box is given as the input. After that, the algorithms are required to track the objective in the later frames on-the-fly. Tracking-by-detection is one of the main research branches of online tracking. However, there still exist two issues in order to improve the performance. 1) The limited processing time requires the model to extract low-dimensional and discriminative features from the training samples. 2) The model is required to be able to balance both the prior and new objectives' appearance information in order to maintain the relocation ability and avoid the drifting problem. In this paper, we propose a real-time tracking algorithm called coupled randomness tracking (CRT) which focuses on dealing with these two issues. One randomness represents random projection, and the other randomness represents online random forests (ORFs). In CRT, the gray-scale feature is compressed by a sparse measurement matrix, and ORFs are used to train the sample sequence online. During the training procedure, we introduce a tree discarding strategy which helps the ORFs to adapt fast appearance changes caused by illumination, occlusion, etc. Our method can constantly adapt to the objective's latest appearance changes while keeping the prior appearance information. The experimental results show that our algorithm performs robustly with many publicly available benchmark videos and outperforms several state-of-the-art algorithms. Additionally, our algorithm can be easily utilized into a parallel program.

DOI： 10.1587/transinf.2014EDP7210

Web of Science

researchmap

記述言語：日本語   掲載種別：研究論文（その他学術会議資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（その他学術会議資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

Many existing image retrieval methods cannot deal with modified query images because modified parts may largely change the original color, shape and texture information. Since every pixel's information of an image has possibility to be changed by editors, local features become less credible. In our method, a very sparse measurement matrix is applied for the compressing procedure. The retrieval database can be compressed effectively by off-line. In order to ensure the retrieval efficiency, this method is not data-driven and does not depend on learning methods. The benchmark includes 5,000 images modified by people according to their preferences. The experimental results show that our method is more efficient than many descriptors on solving this visual similarity measure problem. Even for the images which are modified with complex effects, our method can perform favorably in contrast with other methods. The main contributions of this paper can be summarized as: 1) We propose a globally compressed image descriptor which is simple and effective for solving this visual similarity evaluation problem; 2) We build a benchmark for evaluating the descriptor's performance on solving this problem by comparing with other methods.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

This paper adopts the genetic algorithm optimized flipping scheme to transform the frontal face detector to be profile face detectors for various profile views, rather than the normally building different detectors for different views, or small view ranges. The flipping scheme takes advantage of the over-representation characteristic of the frontal face detector trained with the SURF cascade framework. However, the previous flipping scheme utilizes the classic sliding window method to search for the the hypothetical face axis of symmetry, which is time-consuming. This paper treats this searching as an optimization problem and use the genetic algorithm to optimize the coordinates, the scale and rotation angle of the profile face. We implemented the real time detection of profile faces from different views with genetic algorithm optimized flipping scheme.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

To make Virtual Reality (VR) world closer to reality, it is necessary to acquire much information from the real-world. Especially for simulation of transportation, it is important for VR world to obtain traffic flow in reality. This paper describes an application to compute traffic flow through a sequence of images captured with video camera. In this research, camera is not set above the roadway but beside the roadway.

DOI： 10.1109/CW.2013.32

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

In this paper, we propose a pose estimation method of an object in three-dimensional space without the pre-learning for AR application in the mobile device. One of the implementation methods of augmented reality is using a two-dimensional marker. A two-dimensional marker is not robust in the case of change in complicated three-dimensional rotation. In our work, this problem is solved by a three-dimensional marker using a cube. In the study of a pose estimation of the three-dimensional object, a model-based method using evolutionary method has been proposed. This method cannot estimate all poses, because the rotation angle of the target object is limited. In this paper, we propose the pose estimation method of a cube using a property of cube with evolutionary method. This method is possible to estimate all pose by using property related to rotation of the cube. We evaluated the proposed method by experiments using the static image of object in real-world.

DOI： 10.1109/CW.2013.24

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：日本語   掲載種別：研究論文（研究会，シンポジウム資料等）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：WILEY-BLACKWELL  

The behavior of the particle swarm optimization (PSO) algorithm is analyzed by regarding its dynamics as a system with multiplicative noise and applying control-theoretic analysis methods. In order to evaluate the convergence and diversity of the PSO algorithm, two new measures related to the decay rate and l2 gain of the PSO dynamics are introduced. These measures are characterized by linear matrix inequalities and are therefore efficiently computed by convex optimization tools. Numerical experiments suggest that the measures are effective enough to evaluate the convergence and diversity of the PSO algorithm, which can lead to better understanding of the PSO algorithm from the viewpoints of exploitation and exploration abilities.

DOI： 10.1002/asjc.263

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：SOC INSTRUMENT CONTROL ENGINEERS JAPAN  

In this paper, we propose an effective method for frontal face detection. The proposed method uses a template matching with the genetic algorithm (GA). This method uses a template which shows a distribution of average intensities on a human face. The template and some candidate areas which are searched by GA are compared, and theirs similarities are calculated. The proposed method is robust to geometric transformation of the face. This method can sense the face, such as a location, and an angle of rotation of the face. In some experiments, the proposed method can detect and sense the frontal human face.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：ICIC INTERNATIONAL  

This paper proposes a simultaneous tuning method for PID controllers and dead-zone compensatom based On particle swarm. optimization By the proposed algorithm, PID gains and dead-zone parameters are tuned online to minimize the error between the control output. and the reference input However, since ultrasonic motors intrinsically have a varying dead-zone in the control input associated with applied load torque, dead-zone compensation is required for precise control of these motors The proposed method is applied to an ultrasonic motor, and its effectiveness is verified by experimental results

Web of Science

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

The decline in the quality of patient's safety control is a problem, because the number of caretakers is reduced by the acceleration of demographic aging in an elder care facility. Especially, the detection of getting up from the bed is very important for preventing patients from falling and wandering unbreakable. In our previous study, we have developed the getting up detection device with an ultrasonic radar, which is safe, cheap, and break-proof. However, if there are many patients in a ward, it is difficult to use some ultrasonic radars. The reason is that if some ultrasonic radars, which have the same frequency, are used in same ward, the ultrasonic signals are coherent with each other. To solve this problem, we propose a novel incoherent method. This method is achieved by improving the software in the device at a low cost. © 2010 The Institute of Electrical Engineers of Japan.

DOI： 10.1541/ieejias.130.599

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：WILEY-BLACKWELL  

This paper proposes a PI control method adjusted by a genetic algorithm (GA) for all ultrasonic motor (USM). In the proposed method, proportional and integral gains are adjusted using a GA in order to compensate the characteristic variations of the plant. The effectiveness of the proposed method is confirmed by experiments. (C) 2009 Wiley Periodicals, Inc. Electr Eng Jpn, 169(1): 59-65, 2009; Published online in Wiley InterScience (www. interscience.wiley.com). DOI 10.1002/eej.20842

DOI： 10.1002/eej.20842

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：ICIC INT  

In this paper, we propose a system for automatic detection and recognition of vehicle license plates using template matching, genetic algorithms and neural networks. In the license plate detection part, an artificial template that consists of an outer rectangle that encloses two smaller rectangles is constructed. The size of the outer and inner rectangles, the positioning, size, orientation, position and color of the template are all controlled using the genetic algorithm. To fill the areas outside the inner rectangle, random selection of color pixels from a plate color database are used. In the character recognition part, we combine two methods into a hybrid system. In the first method, we train neural networks to recognize the characters and in the second method, we use template matching. To control the size of both the neural network inputs and the template, we apply a genetic algorithm to guide the search. The final system accuracy achieved is 96.8% for character recognition and 98.01% for license plate detection.

Web of Science

researchmap

記述言語：日本語   出版者・発行元：日本AEM学会  

Ultrasonic motor (USM) employs vibration generated by adding voltage to piezoelectric element as a drive source. In this paper, we propose the control method using variable gain type IMC-PID control combined with neural network (NN) for USM. There are limitations of control performance on the conventional fixed-gain type PID and IMC-PID control because USM causes serious characteristic changes during operation and contains non-linearity caused by frictions. For characteristic changes of the plant, we use a method of variable gain type IMC-PID control. In this method, plant parameters are identified on line and these estimated parameters are used for adjusting three gains of PID. However it is not useful for non-linearity of the plant. Then we propose an improved design scheme of this method combined with NN. NN is used so that the relationship between the input and the output of the plant may become linear.

CiNii Article

CiNii Books

researchmap

その他リンク： https://projects.repo.nii.ac.jp/?action=repository_uri&item_id=310549

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

CiNii Article

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

Stability of the particle swarm optimization algorithm is analyzed without any simplifying assumptions made in the previous works. To evaluate the convergence speed the algorithm, the decay rate is introduced, and a method for finding the largest lower bound of the decay rate is presented. Moreover, it is pointed out that the l(2) gain of the algorithm can be used to measure exploration ability of the algorithm, and a method for finding of the smallest upper bound of the l(2) gain is provided. The above methods are based on linear matrix inequality techniques and therefore are carried out efficiently by using convex optimization tools. Numerical examples are given to show that the analysis methods are reasonable and effective to select the parameters in the algorithm.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

Ultrasonic motor (USM) causes serious characteristic changes during operation and contains non-linearity caused by frictions. To compensate the characteristics, the internal model control - proportional integral differential (IMC-PID) control which requires only one parameter setting beforehand is widely applied to the control of ultrasonic motor. However, there are limitations of control performance on the conventional fixed-gain type IMC-PID) control. So, in this paper we propose an improved method of variable gain type IMC-PID control combined with neural network (NN). In this proposed method, the identification method is adopted to overcome serious characteristic changes and neural network is used to compensate non-linearity. And, the effectiveness of the proposed method has been confirmed by experiments.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：Institute of Electrical Engineers of Japan  

Monetary transactions using Automated Teller Machines (ATMs) have become a normal part of our daily lives. At ATMs, one can withdraw, send or debit money and even update passbooks among many other possible functions. ATMs are turning the banking sector into a ubiquitous service. However while the advantages for the ATM users (financial institution customers) are many the financial institution side faces an uphill task in management and maintaining the cash flow in the ATMs. On one hand, too much money in a rarely used ATM is wasteful, while on the other insufficient amounts would adversely affect the customers and may result in a lost business opportunity for the financial institution. Therefore, in this paper, we propose a daily cash flow estimation system using neural networks that enables better daily forecasting of the money required at the ATMs. The neural network used in this work is a five layered hour glass shaped structure that achieves fast learning even for the time series data for which seasonality and trend feature extraction is difficult. Feature extraction is carried out using the Akamatsu Integral and Differential transforms. This work achieves an average estimation accuracy of 92.6%. © 2009 The Institute of Electrical Engineers of Japan.

DOI： 10.1541/ieejeiss.129.1325

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：Institute of Electrical Engineers of Japan  

The purpose of this study is to develop an interactive interface using eye movement to operate a welfare apparatus, such as a feeding device for an orthopedically-impaired person. A part of the purpose is to eliminate special calibration and re-calibration during the operation. Originalities of this proposed system are as follows: eye sensing with evolutionary processing and interactive operation screen based on some physiological knowledge. The proposed system uses a non-contact type interface by the eye movement. An iris is tracked and eye movement is measured by the evolutionary eye sensing (EES) method. An operation screen is divided into 9 areas, which has visual stimulation using the physiological knowledge. A user can select these areas by eye movement and decide by eye fixation. The effectiveness of the proposed system is evaluated by comparison experiments with 20 subjects. The results indicate that the proposed system is easy to use for first-timer, who can become proficient in operation just after a few exercises. © 2009 The Institute of Electrical Engineers of Japan.

DOI： 10.1541/ieejeiss.129.1288

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：Fuji Technology Press  

In this paper, we propose a fully automatic real time one image face gesture simulation using image morphing. Given a single image of a subject, we create several facial expressions of the face by morphing the image based on prior information stored in a data bank. The process involves the automatic detection of the control points both on the target image and the source data. The source data is a string of frames containing the desired facial expressions. A face detection neural network and a lips contour detector using edges and the SNAKES algorithm are employed to detect the face position and features. Five control points and the lips contour, for both the source and target images, are then extracted based on the facial features. Triangulation method is then used to match and warp the source image to the target image using the control points. In this experiment, using one expressionless face portrait, we create an animation to make it appear like the subject is pronouncing the five Japanese vowels. The final results shows the effectiveness of our method.

DOI： 10.20965/jaciii.2009.p0441

Scopus

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：Institute of Electrical Engineers of Japan  

This paper proposes a scheduled model predictive control method such that input and output constraints are satisfied. In particular, an overshoot constraint as an output constraint is dealt with and an offline technique for finding appropriate feedback gains satisfying the overshoot constraint is presented. Moreover, an online technique for tuning the control input is proposed to improve the output response. The proposed method is applied to a pneumatic servo system, and an experimental result is given to show the effectiveness of the proposed method. © 2009 The Institute of Electrical Engineers of Japan.

DOI： 10.1541/ieejeiss.129.671

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：INT INFORMATION INST  

Eye movement is useful for a no physical contact interface. For practical use, eye tracking should be robust to the low image quality, changes of appearance by eye gaze, and spontaneous or reflex blinking. In this paper, we try to achieve high speed and accuracy eye tracking, which has robustness to these. Moreover, we also try to acquire numerical parameters to represent the eye. We use Evolutionary Video Processing and new fitness function. As a result, the proposed method is robust to the complex changes. The eye tracking accuracy is 96.4%. The processing time is fast enough for real-time processing.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

This paper proposes a decentralized model predictive control method based on a dual decomposition technique. A model predictive control problem for a system with multiple subsystems is formulated as a convex optimization problem. In particular, we deal with the case where the control outputs of the subsystems have coupling constraints represented by linear equalities. A dual decomposition technique is applied to this problem in order to derive the dual problem with decoupled equality constraints. A projected subgradient method is used to solve the dual problem, which leads to a decentralized algorithm. In the algorithm, a small-scale problem is solved at each subsystem, and information exchange is performed in each group consisting of some subsystems. Also, it is shown that the computational complexity in the decentralized algorithm is reduced if the dynamics of the subsystems are all the same. Numerical examples are given to show the effectiveness of the proposed method.

Web of Science

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

In this paper, real-time tracking and information acquisition of lip region of a person during speech in active video scenes are addressed. The lip is deformed by speech, and the size and orientation are changed by camera motion. The difficulty is mainly due to change of these appearances of the lip. We use template matching with a genetic algorithm to overcome these problems. A high speed and accurate tracking scheme using Evolutionary Video Processing is proposed. Usually, a genetic algorithm is unsuitable for a tracking accuracy of 91.6% and an average processing time of 26.0 milliseconds per frame are achieved. © 2008, Taylor &amp
Francis Group, LLC.

DOI： 10.1080/1931308X.2008.10644159

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

Shape memory alloy (SMA) actuators can be deformed by being heated up and cooled down. There are mainly two methods for controlling the deformation of the SMA actuators. The one is to use thermal control devices such as Peltier elements. The other is to electrify the SMA actuators, which utilizes the fact that they are heated up due to joule heat. The goal of this paper is to improve the response speed of the SMA actuators by using the Peltier elements as well as by electrifying the SMA actuators. First, properties of the SMA actuators and the Peltier elements are examined. Next, an online control method is proposed to find appropriate time intervals for heating and cooling the SMA actuators with the Peltier elements. The proposed method is based on a golden section search which is a simple but efficient optimization method. Then a control method for preheating and precooling the SMA actuators is presented. The experimental results are given to show the effectiveness of the proposed method.

Web of Science

researchmap

記述言語：日本語   出版者・発行元：ＩＴヘルスケア学会  

少子高齢化の進展に伴い高齢者の福祉現場では様々な問題が生じている。特に中小規模の高齢者福祉施設や一般家庭では介護者の人手不足が深刻で高齢者の移動を自動的に検知し通報するシステムの開発が望まれている。このためここでは中小規模の高齢者福祉施設や一般家庭でも導入可能であり，安全で，安価で，故障の少ない超音波レーダと電灯線通信を用いた移動検知通報システムを開発したので報告する。本システムは高齢者がベッドから離れようとした場合や居室から出ようとした場合にそれを検知し，管理室や別室の介護者に警報を発する移動検知通報システムである。

DOI： 10.11204/ithc.2.95

CiNii Article

researchmap

出版者・発行元：The Society of Life Support Engineering  

DOI： 10.5136/lifesupport.20.Supplement_53

CiNii Article

researchmap

記述言語：英語   出版者・発行元：日本知能情報ファジィ学会  

In this paper, we propose a fully automatic real time one image face gesture simulation using image morphing. Given a single image of a subject, we create several facial expressions of the face by morphing the image based on prior information stored in a databank. The process involves the automatic detection of the control points both on the target image and the source data. The source data is a string of frames containing the desired facial expressions. A face detection neural network, lips detector and edge detectors are employed to detect the face position and features. Five control points and the lips contour, for both the source and target images, are then extracted bvased on the facial features. Triangulation method is then used to match and warp the source image to the target image using the control points.

DOI： 10.14864/softscis.2008.0.631.0

CiNii Article

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

Ultrasonic motors (USM) causes serious characteristic changes during operation. It is difficult for the conventional internal model control (IMC) proportional integral differential (PID) control to compensate such characteristic changes of the plant. To solve these problems, we propose a method of variable gain type IMC-PID control. In the proposed method plant parameters are identified on line and these estimated parameters are used for adjusting three gains of PID. Then the proposed method makes it possible to compensate characteristic changes of the plant. The effectiveness of the proposed control method have been confirmed by experiments using the existing ultrasonic motors servo system. © 2008 The Institute of Electrical Engineers of Japan.

DOI： 10.1541/ieejias.128.1326

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE COMPUTER SOC  

In this paper, new statistical learning algorithms with kernel function are presented Recently, iterative learning algorithms for obtaining eigenvectors in the principal component analysis (PCA) have been presented in the field of pattern recognition and neural network. However, the Fisher linear discriminant analysis (FLDA) has been used in many fields, especially face image analysis. The drawback of FLDA is a long computational time based on a large-sized covariance matrix and the issue that the within-class covariance matrix is usually singular. In order to overcome this difficulty, we proposed the feature generation method Simple-FLDA which is approximately derived from geometrical interpretation of FLDA. This algorithm is similar to Simple-PCA and does not need matrix operation. In this paper, new statistical kernel based learning algorithms are presented. They are extended versions of Simple-PCA and Simple-FLDA to nonlinear space using the kernel function. Their preliminary simulation results are given for a simple face recognition problem.

DOI： 10.1109/SITIS.2008.52

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

The PID control is widespread most as control methods of an ultra sonic motor (USM). However, setting three gains of the PID control beforehand is not so easy. On the other hand, an internal model control type PID (IMC-PID) control is very effective because it requires only one parameter setting beforehand. However, there are limitations of control performance on the conventional fixed-gain type IMC-PID control because USM causes serious characteristic changes during operation and contains non-linearity caused by frictions. In this paper, we propose the variable gain type IMC-PID combined with neural network (NN). In the proposed method, the identification method is adopted to overcome serious characteristic changes and NN is used to compensate non-linearity. The effectiveness of the proposed method has been confirmed by experiments.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：SPIE-INT SOC OPTICAL ENGINEERING  

In this paper, we have constructed a robot that help people with disabilities of upper extremities and advanced stage amyotrophic lateral sclerosis (ALS) patients to eat with their residual abilities. Especially, many of people suffering from advanced stage ALS of the use a pacemaker. And they need to avoid electromagnetic waves. Therefore we adopt ultra sonic motor that don't generate electromagnetic waves as driving sources. Additionally we approach the problem of the conventional meal assistance robot. Moreover, we introduce the interface with eye movement so that extremities can also use our system. User operates our robot not with hands or foot but with eye movement.

DOI： 10.1117/12.783967

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

In this paper, we propose a new statistical learning algorithm. This study quantitatively verifies the effectiveness of its feature extraction performance for face information processing. Simple-FLDA is an algorithm based on a geometrical analysis of the Fisher linear discriminant analysis. As a high-speed feature extraction method, the present algorithm in this paper is the improved version of Simple-FLDA. First of all, the approximated principal component analysis (learning by Simple-PCA) that uses the mean vector of each class is calculated. Next, in order to adjust within-class variance in each class to 0, the vectors in the class are removed. By this processing, it becomes high-speed feature extraction method than Simple-FLDA. The effectiveness is verified by computer simulations using face images.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

Purpose of our works is development of an interface by eye movement for physically impaired people due to spinal cord injury and neurological disease. Eye appearance is changed by eye blinking, head motion, eye movement, and individual difference. Moreover, the system must track the eye and acquire its information simultaneously. For these problems, we use Genetic Eye Tracking method with an artificial iris template. In this paper, we propose an application of the Genetic Eye Tracking, and the effectiveness is evaluated by simulations. As the result, this proposed system achieves high accuracy.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

We propose the detection and reporting device for confirming the safety of the patients at small nursing facilities or ordinary houses. The device combines the ultrasonic sensor which detects patients' getting out of bed with power line communication (PLC) as a reporting system. The ultrasonic sensor in this device detects elderly person's getting out of bed and lying down to bed. The effectiveness of the proposed device has been confirmed by the field experiments in welfare facilities for elderly people.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：WORLD SCIENTIFIC AND ENGINEERING ACAD AND SOC  

In this paper, a real time face detection method using several small size neural networks and a genetic algorithm with adaptive search area control is proposed. Neural networks and genetic algorithms may not be suitable for real time application because of their long processing times. However, in this paper, we show how fast speeds can be achieved using small effective neural networks and a genetic algorithm with a small population size that requires few generations to converge. We subdivide the face into several regions, each connected to an individual neural network. This guarantees small size networks and also offers the ability to learn different face regions features using different coding methods. The genetic algorithm is used during the real time search. It extracts possible faces from face candidates that are then tested using the neural networks. The face candidate area is then adaptively reduced depending on the location of the top six face samples. We then performed real time simulation using an inexpensive USB camera to prove the effectiveness of our proposal. We achieved between 98 and 96% accuracy for one or multiple faces respectively at 15 to 8 frames per second.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

We propose the detection and reporting device for confirming the safety of the patients at small nursing facilities or ordinary houses. The device combines the ultrasonic sensor which detects patients' getting out of bed with power line communication (PLC) as a reporting system. The ultrasonic sensor in this device detects elderly person's getting out of bed and lying down to bed. The effectiveness of the proposed device has been confirmed by the field experiments in welfare facilities for elderly people. © 2007 SICE.

DOI： 10.1109/SICE.2007.4421032

Scopus

researchmap

記述言語：日本語   出版者・発行元：The Institute of Electrical Engineers of Japan  

This paper proposes a PI control method adjusted by a genetic algorithm (GA) for an ultra-sonic motor (USM). In the proposed method, a proportional and an integral gains are adjusted using GA in order to compensate characteristic variations of the plant. The effectiveness of the proposed method is confirmed by experiments.

DOI： 10.1541/ieejias.127.663

CiNii Article

CiNii Books

researchmap

その他リンク： https://jlc.jst.go.jp/DN/JALC/00295121381?from=CiNii

記述言語：英語   掲載種別：研究論文（学術雑誌）  

DOI： 10.20965/jaciii.2007.p1030

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

Pneumatic systems intrinsically have hard limits on the control input and output. To cope with these constraints, a scheduled model predictive control method is modified and applied to a pneumatic system. Unlike standard model predictive control methods, this method can deal with systems that have relatively fast dynamics because the online computational complexity is reduced by computing a sequence of feedback gains in advance. In this control scheme, effective offiine and online algorithms for choosing control parameters are presented. Experimental results show the effectiveness of the proposed control method.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：SPRINGER-VERLAG BERLIN  

This paper presents an improved statistical learning algorithm for feature generation in pattern recognition and signal processing. It is approximately derived from geometrical interpretation of the Fisher linear discriminant analysis (FLDA). The principal component analysis (PCA) is popular for data compression and feature extraction. Furthermore, iterative learning algorithms for obtaining eigenvectors in PCA have been presented in such fields. Their effectiveness has been demonstrated in many applications. However, recently FLDA has been often used in many fields, especially face image recognition. The drawback of FLDA is a long computational time based on a large-sized covariance matrix and the issue that the within-class covariance matrix is usually singular. Generally, in FLDA, the inverse matrix of the within-class covariance matrix cannot be obtained, since data dimension is generally higher than the number of data and then it includes many zero eigenvalues. In order to overcome this difficulty, a new iterative feature generation method, the simple-FLDA was proposed by authors. In this paper, further improvement is introduced into the simple-FLDA and its effectiveness is demonstrated for preliminary personal identification problem.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

An estimation method of monotonic signals was proposed lately. This method is quite effective when a priori knowledge of monotonicity of the estimated signals can be used. However, the estimation method was presented only for the case where an estimated signal is scalar and the measured signal is supposed to be generated by an FIR model. This paper extends the estimation method to the cases of multiple signals and IIR models. Numerical examples are provided to illustrate the effectiveness of the extended method.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

researchmap

出版者・発行元：システム制御情報学会  

本稿では，大きさと姿勢に不変な高速な目の追跡を提案する．さらに，位置，大きさ，回転角度などの目を表す数値的なパラメータを得ることを目的とする．提案手法はテンプレートマッチングに基づいており，遺伝的アルゴリズムを用いる．システムの汎用性を考え，テンプレートマッチングに使用されるテンプレート画像は，すべてのユーザに対し，1枚のテンプレートが用いられる．探索率は74%で処理時間は約0.03秒であった．

DOI： 10.11509/sci.SCI06.0.298.0

CiNii Article

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：ACTA PRESS ANAHEIM  

in Japan, diet-linked diseases are increasing owing to the reasons of diversified and meat-oriented diet. Disease of an internal organ and brain disease are those instances. Furthermore regardless of race, increase in fatalities by cancer, cardiac disease and cerebropathy is a social issue. Medical equipments as the CT, MRI and ultrasonography (US) is used for these sick discoveries. The demand for Medical Equipments has increased. The doctor's load is increasing as it increases. The purpose of this work is the construction of an automatic diagnosis support system for CT images in order to reduce the doctor's load. Toward this end, in this paper, a method to extract liver tumors in CT images using a real-coded genetic algorithm is proposed. Conventionally, a threshold is necessary to extract an object from an image. However, such a method is not effective for CT images because gray scale values are different in each image. Therefore, in this paper, we propose the method for extracting the tumor in the liver from the CT image without the need of a threshold. In this method, a polygon enclosure of the liver tumor is extracted using a GA.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：ACTA PRESS ANAHEIM  

We propose a high-speed size and orientation invariant eye detection in an active scene. As part of the objectives, we also try to acquire numerical parameters to represent the eye, for example, position, scaling and rotation angle. The information is useful for many applications, where high performance is required, such as eye gaze detection or estimation for robot perception and mobile devices interfaces. The difficulty in eye tracking is mainly due to motion of the human head and the active scene by free camera motion. The proposed system is based on template matching using genetic algorithms (GAs). The generality of the proposed method is provided by the artificial iris template used. Various geometric changes of iris can be supported by GAs. In this paper, eye detection method with GAs is proposed. Moreover the effectiveness is evaluated by comparison with a representative method. These results indicate that the detection accuracy of the both methods is same (about 90%) in the simple situation. On the other hand, the proposed method is more robust to the complex situation than the a representative method. The processing time is fast enough for real-time processing.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：SPRINGER-VERLAG BERLIN  

In this paper, a simple 3D edge based template for pose invariant face detection creation using side and front profiles of a general face is proposed. The 3D template is created using the edges that are always most likely to be extracted from a face given a certain pose. Bezier curves are used to create the template. When testing the template, genetic algorithms are used to guide the matching process thereby greatly reducing the total computation time required. The genetic algorithm automatically calculates the angle and the size of the template during the matching. An average pose invariant face detection accuracy of 84.6% was achieved.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：日本知能情報ファジィ学会  

In this paper, a method of downsizing a genetic algorithm (GA) for lips detection is described. As part of the objectives, we also try to acquire information to represent the lips. GA is very useful for optimization of parameters in complex image matching. When GA is applied to the real-time processing, the search speed is a issue with keeping the accuracy. Using small population is one of the approaches to improve the GA search speed. However, this can cause the low accuracy, because diversity of the population will be lost. In order to avoid this problem, we propose a downsized GA with automatic search domain control. In this paper, the optimization problem is template matching. The target image includes a face of talking person and has significant changes of a whole scene by camera motion. The template image is only one closed mouth and prepared for each environment and person, because of personal use. In simulations, a simple image is used, which is the basis of video sequence. Our proposed method is verified by comparison of a process of evolution between a downsized GA and a standard GA.

DOI： 10.14864/softscis.2006.0.1312.0

CiNii Article

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：INT INST INFORMATICS & SYSTEMICS  

We propose a high-speed size and orientation invariant eye tracking method, which can acquire numerical parameters to represent the size and orientation of the eye. In this paper, we discuss that high tolerance in human head movement and real-time processing that are needed for many applications, such as eye gaze tracking. The generality of the method is also important. We use template matching with genetic algorithm, in order to overcome these problems. A high speed and accuracy tracking scheme using Evolutionary Video Processing is proposed. Usually, a genetic algorithm is unsuitable for a real-time processing, however, we achieved real-time processing. The generality of this proposed method is provided by the artificial iris template used. In our simulations, an eye tracking accuracy is 97.3% and, an average processing time of 28 milliseconds per frame.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   掲載種別：研究論文（学術雑誌）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）  

In this paper, a lips extraction method that can extract lips region from varying lips shape at the moment of speech by using only one template image is described. The method that is proposed in this paper, has invariance for an open and closed mouth, showing or not showing any teeth, and has high speed and high extraction accuracy in consideration for characteristics of the lips by using a genetic method. This method uses the template matching using a genetic algorithm. Furthermore, color of lips and characteristics of the lips shape variances at the moment of speech in this system are utilized. The effectiveness of this method is demonstrated with only one template for each person being tested and a search object, that is, the varying lips shape at the moment of speech of vowels by means of computer simulations. These computer simulations indicate that this method can extract the varying lips shape at the moment of speech by using only one template. Moreover, in the extraction processing of every vowel, a igh speed and high extraction accuracy can be obtained. © 2004, The Institute of Electrical Engineers of Japan. All rights reserved.

DOI： 10.1541/ieejeiss.124.128

Scopus

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

In this paper, real-time detection and tracking of lips region of a talking person in natural scenes is addressed. In particular, we try to acquire numerical parameters to represent the lips information. Because, this information is very important for many applications, such as audio-visual speech recognition, robot perception, and interface of mobile devices. The difficulty lies in deformations and geometric change of lips, by speech and free camera work. Our proposed system is based on template matching with genetic algorithms (GAs). In our previous system, there is A trade-off between accuracy and a processing time. However, we can overcome this by two new methods: (a) a flexible control of a search domain, (b) inheritance of genetic information between video frames. We demonstrated the, effectiveness of our proposed system by using some 5 seconds video sequences. The average results are that the accuracy is 94.44% and the processing time is 4.50 seconds.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：SPRINGER-VERLAG BERLIN  

In this paper, a lips extraction method by genetic algorithms (GAs), which has simple, high speed and high accuracy properties is described. This method is a visual front end of audio-visual speech recognition systems for mobile devices. Our method is based on template matching with GA. GA can use a significant amount of computation time, because the calculation and test of fitness values for generation x population is necessary. Furthermore GA is basically a global optimization algorithm. So, premature convergence in GA often causes it to fail to reach the global optimum. To extract lips region at high speed with high accuracy, the population must be small and local search must be more efficient. In order to overcome these problems, we propose the method to control search domain flexibly. We compare this method with the previous method by small population. The experimental results show the proposed method is effective.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：ACTA PRESS  

In this paper, as a new visual front end of audio-visual speech recognition systems, a simple method for lips region extraction is proposed. This method has robustness for varying shape and three-dimensional geometric changes by using only one image template. The image data is represented by x component (redness) of the Yxy colour space. The system environment is that a camera and human head run around separately for mobile devices. In our previous study, we tried to extract lips region, which has varying shape by speech, and the geometric changes of lips on two-dimension. Then, we extend this previous method to three-dimension by using projective geometry. In our simulations, we try to extract lips region from a face image with back-round, varying shape by speech, geometric changes and face horizontal direction change. By this simulation, we can obtain high extraction accuracy.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（学術雑誌）   出版者・発行元：SPRINGER-VERLAG BERLIN  

A method to track a moving object that has a continuous shape deformations is described in this paper. Moreover this method aims to also obtain information about the shape deformations. This method uses template matching, where the template is deformed. The deformation of the template and matching process are performed using a genetic algorithm, Our conventional method does not have a good search efficiency. Therefore, a new method, which distinguishes between an object function and fitness function, is proposed in this paper. Comparisons between our conventional method and the new method are also simulated. As an object for this simulation, a lips region during speech is used. From results of this simulation, it comes to light that the proposed method is more better than the conventional one in tracking accuracy and speed.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）   出版者・発行元：IEEE  

In this paper, a lips extraction method that can extract lips region from varying lips shape at the moment of speech by using only one template image is described. The method that is proposed in this paper, has invariance for an open and closed mouth, showing or not showing any teeth, and has high speed and high extraction accuracy in consideration for characteristics of the lips by using a genetic method. This method uses the template matching using a genetic algorithm. Furthermore, colour of lips and characteristics of the lips shape variances at the moment of speech in this system are utilized. The effectiveness of this method is demonstrated with only one template for each person being tested and a search object, that is, the varying lips shape at the moment of speech of vowels by means of computer simulations. These computer simulations indicate that this method can extract the varying lips shape at the moment of speech by using only one template. Moreover, in the extraction processing of every vowel, a high speed and high extraction accuracy can be obtained.

Web of Science

researchmap

記述言語：英語   掲載種別：研究論文（国際会議プロシーディングス）  

researchmap

記述言語：日本語   出版者・発行元：一般社団法人電子情報通信学会  

本橋ではシンプルかつ,高速な口唇領域の抽出法について述べる.この手法はモバイル機器を対象とした視聴覚情報を用いた音声認識の画像的前処理として使用することができる.探索方法としてマッチング処理に遺伝的アルゴリズム(GA)を用いたテンプレートマッチングを使用する.通常,GAは計算負荷が高い.これは個体数×世代数分の評価を必要とするからである.また,GAは広域探索手法である.我々の過去の研究から,失敗する大きな原因として,GAの早熟な収束による広域解への未到達が考えられる.更に,高速な処理実現するには,より少ない個体数で局所探索の効率を向上させる必要がある.これらの問題を克服するために,探索領域を制御する手法を提案する.少ない個体数を用い,探索領域の制御を行う場合と通常の場合で比較実験を行った.結果から本提案手法が有効であることが示された.

CiNii Article

CiNii Books

researchmap

記述言語：日本語   出版者・発行元：一般社団法人電子情報通信学会  

CiNii Article

CiNii Books

researchmap

受賞区分：国内学会・会議・シンポジウム等の賞  受賞国：日本国

researchmap

受賞区分：学会誌・学術雑誌による顕彰  受賞国：日本国

researchmap

受賞区分：学会誌・学術雑誌による顕彰  受賞国：日本国

researchmap

受賞区分：出版社・新聞社・財団等の賞  受賞国：日本国

researchmap

受賞区分：国際学会・会議・シンポジウム等の賞  受賞国：日本国

researchmap

受賞区分：国際学会・会議・シンポジウム等の賞  受賞国：アメリカ合衆国

researchmap

受賞区分：出版社・新聞社・財団等の賞  受賞国：日本国

researchmap

受賞区分：出版社・新聞社・財団等の賞  受賞国：日本国

researchmap