研究者詳細 - 吉田　香

論文 - 吉田　香

分割表示125 件中 21 - 40 件目／全件表示 >>

Simple Meta-optimization of the Feature MFCC for Public Emotional Datasets Classification 査読有り国際誌

de la Cal E., Gallucci A., Villar J.R., Yoshida K., Koeppen M.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 12886 LNAI 659 - 670 2021年01月

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）

A Speech Emotion Recognition (SER) system can be defined as a collection of methodologies that process and classify speech signals to detect emotions embedded in them [2]. Among the most critical issues to consider in an SER system are: i) definition of the kind of emotions to classify, ii) look for suitable datasets, iii) selection of the proper input features and iv) optimisation of the convenient features. This work will consider four of the well-known dataset in the literature: EmoDB, TESS, SAVEE and RAVDSS. Thus, this study focuses on designing a low-power SER algorithm based on combining one prosodic feature with six spectral features to capture the rhythm and frequency. The proposal compares eleven low-power Classical classification Machine Learning techniques (CML), where the main novelty is optimising the two main parameters of the MFCC spectral feature through the meta-heuristic technique SA: the n_mfcc and the hop_length. The resulting algorithm could be deployed on low-cost embedded systems with limited computational power like a smart speaker. In addition, the proposed SER algorithm will be validated for four well-known SER datasets. The obtained models for the eleven CML techniques with the optimised MFCC features outperforms clearly (more than a 10%) the baseline models obtained with the not-optimised MFCC for the studied datasets.

DOI： 10.1007/978-3-030-86271-8_55

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85115856860&origin=inward

Hand Position Tracking Correction of Leap Motion Controller Attached to the Virtual Reality Headset 査読有り

BELANOVÁ Dorota, YOSHIDA Kaori

International Journal of Biomedical Soft Computing and Human Sciences: the official journal of the Biomedical Fuzzy Systems Association ( バイオメディカル・ファジィ・システム学会 ) 25 ( 1 ) 29 - 37 2020年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

Considering the increasing number of patients with disabilities, especially of upper extremity, we decided to develop hand rehabilitation virtual reality system for post-stroke patients using leap motion controller. However, as the device has not met the required accuracy of hand position tracking, we created a real-time low-computational cost correction algorithm lessening the problem. This paper outlines the challenges of hand position tracking accuracy identifying the parts contributing to the overall error, proposes a method of correction suitable for each of the parts separately and brings the parts together into an algorithm maximizing the hand position correction. The algorithm was tested on twelve healthy participants in two series of four tasks. Results showed our algorithm improved the hand position accuracy considerably in all the cases.

DOI： 10.24466/ijbschs.25.1_29

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130007998938

The Concept of Intelligent Space with a Robot 査読有り

BELANOVÁ Dorota, MACH Marián, SINČÁK Peter, YOSHIDA Kaori

International Symposium on Affective Science and Engineering 2019 ( 0 ) 1 - 4 2019年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

Intelligent space is where people may easily interact with computers and robots, changing the environmental conditions based on users' needs to create more comfortable and satisfying surroundings. In our proposed system, we would integrate an intelligent space with a robot that would act as an active part of the environment. The system is illustrated in a few caregiver scenarios, showing the actions of a robot. The skills of a robot are explained, presenting the advantages and contributions of the system.

DOI： 10.5057/isase.2019-C000024

Kyutacar

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130007656892

Tensor self-organizing map for kansei analysis 査読有り

Itonaga K., Yoshida K., Furukawa T.

Proceedings - 2018 Joint 10th International Conference on Soft Computing and Intelligent Systems and 19th International Symposium on Advanced Intelligent Systems, SCIS-ISIS 2018 796 - 801 2018年12月

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）

© 2018 IEEE. In Kansei analysis, impressions of various objects are commonly measured using evaluation words. When using this approach, it is necessary to examine all combinations of three elements: subjects, objects, and evaluation words. However, the exhaustive analysis required is not an easy task because of the enormous number of combinations. Additionally, if it is necessary to reveal the relationship between the impressions and physical features of objects such as colors or shapes, the number of combinations increases enormously and the task becomes unrealistic. In this paper, we introduce a method called the tensor self-organizing map (TSOM) that visualizes the relationships between the elements. We applied the TSOM to Kansei analysis of landscape images and studied how the impressions were dependent on the subjects. We also investigated the relationships between these subject-dependent impressions and the physical features. Through these experiments, we demonstrate that the TSOM can be a useful tool for Kansei analysis.

DOI： 10.1109/SCIS-ISIS.2018.00132

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85067131361&origin=inward

An Efficient Replay Method to Prevent SURF-Based Face Detection 査読有り

Dadkhah S., Koppen M., Sadeghi S., Yoshida K.

Proceedings - 2018 5th Asia-Pacific World Congress on Computer Science and Engineering, APWC on CSE 2018 18 - 25 2018年12月

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）

© 2018 IEEE. During the past couple of years, a considerable number of digital images are gathered by search engines, or uploaded by users in different social websites. These images could contain a variety of variables such as different faces, scenes, and unusual objects. The advancement of digital face recognition algorithms is sensational, recognition algorithms used by some of the social websites are even stronger than most secret service establishments. However, there is always a concern about how to protect the user's privacy when they do not want to be recognized automatically by these services. In this research, a SURF-based face recognition algorithm along with a featurebased attack to mislead this type of scheme is proposed. The recognition algorithm use different machine learning technique along with second layer descriptor counting function which makes it more powerful. However the primary focus of this research is to provide a method which can mislead this type of recognition systems with a smallest possible degradation to the image representation. The experimental results show that the proposed plan is successful to fool the SURF algorithm and its recognition system by 70%.

DOI： 10.1109/APWConCSE.2018.00012

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85074295713&origin=inward

Updatable indoor localization based on BLE signal fingerprint 査読有り

Benaissa B., Yoshida K., Koppen M., Hendrichovsky F.

Proceedings of the 2018 International Conference on Applied Smart Systems, ICASS 2018 2018年11月

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）

© 2018 IEEE. The need for a robust indoor localization solution is mounting in the recent few year, as the IoT technology is growing fast. In this paper, we discuss existing indoor localization approaches, focusing on the signal strength fingerprint approaches, to which belongs the presented study and based on Bluetooth Low Energy signal. The proposed approach employs the Radial Basis Functions (RBF) to create model that describe the real signal strength-position relationship. The present method is fully implemented on a smart phone, namely, offline data collection, model computation and online position estimation. This allow updating the model by collecting more signal data, therefore improving the localization accuracy with small efforts. The presented results shows that updating the model effectively improves the accuracy of estimation.

DOI： 10.1109/ICASS.2018.8652035

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85063516885&origin=inward

上半身動作を用いたロボットの回避方向意図の伝達査読有り

石川友紀也, 末松圭史, 大坪正和, 吉田香

日本知能情報ファジィ学会論文誌 30 ( 5 ) 709 - 716 2018年10月

Path planning on robot based on D∗ lite algorithm 査読有り

Belanova D., Mach M., Sinčák P., Yoshida K.

DISA 2018 - IEEE World Symposium on Digital Intelligence for Systems and Machines, Proceedings 125 - 130 2018年10月

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）

© 2018 IEEE. The increasing need of autonomous behavior of robots in fields of science and technology formed the requirement for path planning implemented by the robot without the human assistance. In this paper, D∗ Lite, which is a path planning graph-based algorithm, was used in order to compute the shortest path from a start to goal point in a real environment and make a Pepper robot move in a computed trajectory. The movement of robot was conducted in a static environment, with the map of the environment already known. This paper is a first step in the research focusing on a creation of a so-called intelligent workspace.

DOI： 10.1109/DISA.2018.8490605

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85056782104&origin=inward

An efficient ward-based copy-move forgery detection method for digital image forensic 査読有り

Dadkhah S., Koppen M., Sadeghi S., Yoshida K., Jalab H., Manaf A.

International Conference Image and Vision Computing New Zealand 2017-December 1 - 6 2018年07月

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）

© 2017 IEEE. In recent years, with the improvement of digital manipulation tools, image forgeries have become a genuine social issue. Copy-move forgery is an important type of digital manipulation which is done by duplicating some parts of the image with the purpose of hiding specific information. Verifying the authenticity of the digital image can become vital for different forensic organizations. In this paper, an efficient copy-move detection algorithm with three level of Ward linkage based clustering is proposed. The proposed algorithm extracts the local image features with the help of Scale-invariant feature transform algorithm (SIFT). In the proposed method, a Ward-based clustering scheme followed by three levels of false-descriptor elimination is applied. The proposed elimination system utilizes Euclidean distance to improve the accuracy of the region detection scheme. The efficiency of the proposed algorithm is examined against several different post-processing attacks such as multilevel forgeries, noises, and JPEG compression. Experimental results illustrate that the proposed algorithm improves the detection accuracy by eliminating false descriptors.

DOI： 10.1109/IVCNZ.2017.8402472

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85050008851&origin=inward

Phone Application for Indoor Localization Based on Ble Signal Fingerprint 査読有り

Benaissa B., Hendrichovsky F., Yishida K., Köppen M., Sinčák P.

2018 9th IFIP International Conference on New Technologies, Mobility and Security, NTMS 2018 - Proceedings 2018-January 1 - 5 2018年03月

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）

© 2018 IEEE. Indoor position of devices and persons is a very valuable information for context-aware devices and human activity recognition. The need of such information grows bigger by the advancement of IoT technologies. In this paper, we discuss existing indoor positioning approaches and their applications, give specific interest to the signal strength fingerprint approaches, to which belongs the presented paper. Bluetooth Low Energy Beacons are used as the source of the signal, and the Radial Basis Functions method is employed to create a model that relates to signal strength and position. The proposed approach allows the possibility of making all the computation on a smartphone application, namely, offline data collection and model computation, and online position estimation. Besides, it allows updating the model later.

DOI： 10.1109/NTMS.2018.8328729

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85047863726&origin=inward

Bad Ai: Investigating the Effect of Half-Toning Techniques on Unwanted Face Detection Systems 査読有り

Dadkhah S., Köeppen M., Sadeghi S., Yoshida K.

2018 9th IFIP International Conference on New Technologies, Mobility and Security, NTMS 2018 - Proceedings 2018-January 1 - 5 2018年03月

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）

© 2018 IEEE. Currently, automatic face detection systems are widely used in various social media. The improvement of the automated intelligent face recognition can easily result in the invasion of the user 's privacy if the predefined functions for their so- called artificial intelligent systems are designed with no respect to the user privacy, this means that users do not get the right to prevent these types of automatic face detection. The problem does not stop here, and recently some organization attempted to use automated face recognition system as a digital personal judgment tools. For instance, judging if a person is criminal only based on the attributes of his/her face. In the present technology, spreading the personal information such as individual digital images are fast and inevitable. Thus, in this paper, some of the techniques to evade automatic face recognition is investigated. The focused of this article is to highlight the advantages of different Half-Toning algorithms concerning avoiding unwanted automated face detection/recognition. In the experimental phase, the result of varying filtering and modification algorithms compared to the proposed filtering technique.

DOI： 10.1109/NTMS.2018.8328726

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85050981631&origin=inward

BAD AI - Investigating the Effect of Half-Toning Techniques on Unwanted Face Detection Systems 査読有り

Sajjad Dadkhah and Mario Koeppen, Somayeh Sadeghi, Kaori Yoshida

the 9th IFIP International Conference on New Technologies, Mobility & Security 2018年02月

Phone application for indoor localization based on BLE signal mapping 査読有り

Filip Hendrichovsky, Brahim Benaissa, Peter Sinčák, Kaori Yoshida

MobiCASE Student Workshop (MCSW2018) in conjunction with the 9th International Conference on Mobile Computing, Applications and Services (MobiCASE 2018) 2018年02月

ユーザモデリングにおける課題と今後の展望査読有り

吉田香

知能と情報 ( 日本知能情報ファジィ学会 ) 30 ( 4 ) 219 - 221 2018年01月

Interactive Evolutionary Computation of Color Palette Design Enhanced by Impression Words 査読有り

Makiwan D., Yoshida K., Köppen M.

2017 International Conference on Platform Technology and Service, PlatCon 2017 - Proceedings 2017年03月

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）

© 2017 IEEE. Color is a powerful tool to describe and interpret information from an image. Colors can be stored in an array called color palette. In the world of product design, palettes are mostly used to inspire creativeness and Interactive Genetic Algorithm has been widely increased to develop wide creativity solutions and generate new ideas for colors. When it comes to the choice of colors, the users do not normally have a clear idea about it. What users can do is use the color palette that is available as the standard palette and produce a color combination of their own choice. Thereafter, from the creation of a new palette, impression words also be shown to the user appeared to represent matching ideas or feelings from the user about the new palette. The aim of this research is to achieve user satisfaction of color palette and try to assist in the selection process by directing the process of optimizing the color combination. This solution can improve the interface designs in term of enhancing human perception skill.

DOI： 10.1109/PlatCon.2017.7883712

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85018164768&origin=inward

CoWoOZ - A cloud-based teleoperation platform for social robotics 査読有り

Magyar G., Sinčák P., Magyar J., Yoshida K., Manzi A., Cavallo F.

SAMI 2017 - IEEE 15th International Symposium on Applied Machine Intelligence and Informatics, Proceedings 49 - 54 2017年03月

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）

© 2017 IEEE. This paper presents our cloud-based teleoperation platform, called CoWoOZ - ClOud-based Wizard of OZ. It describes similar systems, such as Robot Management System (RMS), OpenWoZ, CoWoOZ's predecessor Telescope and the first version of the discussed platform. A special emphasis is given to the description of new functionalities added to the system, e.g. scenario management, custom script builder, client application, etc. The paper also contains a comparison of these systems, which can help researchers to choose the right one for their work.

DOI： 10.1109/SAMI.2017.7880361

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85017665683&origin=inward

Activity and Emotion Recognition for Elderly Health Monitoring 査読有り

BENAISSA Brahim, KÖPPEN Mario, YOSHIDA Kaori

International Journal of Affective Engineering 17 ( 2 ) 81 - 88 2017年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

The problems facing elderly who are living independently, are considered to be one of the most important motivations of the activity recognition research. The advances in sensing technologies allow collecting several types of data and communicate it wirelessly. However, most existing activity recognition systems requires pre-calculated pattern recognition models. This paper lays the theoretical foundations of a real-time methodology for activity and emotional recognition based on body and environment sensors simultaneously, then tackles one aspect of the method which is the path estimation using chest-mounted IMU sensor, for which a zero velocity update criteria is proposed. Finally path estimation results for sitting down, laying down, falling down and standing up are discussed.

DOI： 10.5057/ijae.IJAE-D-17-00020

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130006728077

Tensor SOMによる感性の評価者・評価対象・評価語の同時分析査読有り

糸永恭平, 岩崎亘, 上村洋平, 吉田香, 古川徹生

知能と情報 29 ( 6 ) 661 - 669 2017年01月

記述言語：日本語掲載種別：研究論文（学術雑誌）

本研究の目的は，感性評価における評価者と評価対象，および評価語の三者の関係を同時分析する解析法の開発である．われわれはTensor SOMを用いて評価者マップ，評価対象マップ，評価語マップの3つを同時生成するとともに，マップを彩色することで三者の関係を可視化した．本手法で風景画像集合を解析したところ，各画像がどのような印象を持つか，また評価者によって印象がどのように変わるかを可視化できた．さらに本手法の産業への実応用をめざして解析ツールを開発し，自動車のインストルメントパネルデザインに関する感性・嗜好解析を行った．

DOI： 10.3156/jsoft.29.6_661

CiNii Article

その他リンク： https://ci.nii.ac.jp/naid/130006250681

The price of unfairness 査読有り

Koeppen M., Yoshida K.

Proceedings - 2016 International Conference on Intelligent Networking and Collaborative Systems, IEEE INCoS 2016 463 - 468 2016年10月

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）

© 2016 IEEE.Here, measures for unfairness in resource allocation problems are studied. The extremity of such measures should match with an intuitive concept of an unfair allocation. To achieve a computational model for unfairness, various fairness models are extended to cover a related meaning of unfairness as well. In addition to the lexicographic maxmax relation, a model based on exponential utilities is introduced. Furthermore, a new mean, the symmetric Lehmer mean is identified as being able to favour allocations restricted to a subset of users only. Case examples of network flow control problems show feasibility of the proposed unfairness measures, as well as specific differences. Especially the symmetric Lehmer mean appears capable to handle unfairness in a much more nuanced way, while lexicographic maxmax appears as the computationally most convenient and also most intuitive measure. In all cases it can be shown that unfair allocations are not always efficient and that there is a price of unfairness as well.

DOI： 10.1109/INCoS.2016.71

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84998694468&origin=inward

Comparison of Near and Far Infrared Vision for Biometric Applications 査読有り

Mazandarani E., Lucke U., Yoshida K., Köppen M.

Proceedings - 2015 International Conference on Computer Application Technologies, CCATS 2015 34 - 37 2016年01月

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）

© 2015 IEEE.In this paper, we provide the idea of analyzing the infrared vein patterns in the back of the hand combined with the appendant optical picture of the back of the hand for biometric purposes. We present results of an experiment which will demonstrate differences between near infrared and far infrared images as initial images for creating templates for biometrical purposes. Near infrared images offer a good quality for extracting hand veins on the one hand but contain no thermographic information as far infrared images do. The idea of this paper is to make counterfeit more difficult by using the infrared image with thermographic information and the appendant optical image with spatial information. That is why only far infrared images are useable in this case. This Paper will show a possibility to create a biometrical template based on far infrared images.

DOI： 10.1109/CCATS.2015.17

Scopus

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=84964746678&origin=inward

　前のページ - 次のページ　

<吉田 香>

論文 - 吉田 香

<吉田　香>

論文 - 吉田　香