研究者詳細 - 吉田　香

The Impact of Large Sample Datasets on Hand Gesture Recognition by Hand Landmark Classification 査読有り国際誌

AHMAD Khawaritzmi Abdallah, SILPANI Dian Christy, YOSHIDA Kaori

International Journal of Affective Engineering ( 日本感性工学会 ) 22 ( 3 ) 253 - 259 2023年07月

担当区分：最終著者記述言語：英語掲載種別：研究論文（学術雑誌）

Today, almost all aspects of life are influenced by technology to simplify these aspects of life. Researching how humans and computers interact is very important to build a good model for better human-computer interaction in the future. Using technology such as computer vision, we can now collect the information we are looking for directly from a human. Humans can use many kinds of modalities to interact with computers. Hands are perhaps the largest source of body language information after the face. To understand the gesture’s meaning, we can use MediaPipe Hands, developed at Google LLC, as a method to track and recognize human hands. However, if we want to understand some kinds of hand gestures using MediaPipe Hands, we need to create a condition using if-else manners. This research tried to collect the wide variety of each hand gesture using the 21 key points in x, y, and z coordinates as a feature. We chose Support Vector Machine (SVM) and Artificial Neural Network (ANN) as classifiers to validate the impact of large sample datasets. This research found that ANN is the best among all the methods we used as a classifier method for the 3D value of 21 key points from the hand skeleton. The accuracy and F1-score from ANN were 98.4% accuracy and 98.2% F1-Score, respectively, representing the best performance for each class of all the methods we used in this research.

DOI： 10.5057/ijae.ijae-d-22-00022

その他リンク： https://www.jstage.jst.go.jp/article/ijae/22/3/22_IJAE-D-22-00022/_pdf

STATIC HAND GESTURE RECOGNITION BASED ON MULTI-FEATURE EXTRACTION WITH CONTEXT FOR USER INTENSION 査読有り

Ahmad K.A., Silpani D.C., Yoshida K.

ICIC Express Letters, Part B: Applications ( ICIC International 学会 ) 15 ( 8 ) 779 - 786 2024年08月

担当区分：最終著者記述言語：英語掲載種別：研究論文（学術雑誌）

Hand gesture is a source of information that contains much information, which has the advantage that this modality does not contain private information as much as information obtained from facial modalities. However, obtaining information from hand gestures still encounters several challenges, one of which is the diversity of gestures within one target class which is influenced by many factors such as the culture and personality of the individual. In this study, we propose a method of combining hand gesture features in the form of keypoints features, convolutional neural network features, and manually generating captions as contexts to improve model performance so that the model can classify hand gestures based on the intent of the hand gesture with context support. As a result, the CNN model, which initially experienced over-fitting when trying to classify hand gestures based on the intent of the hand gesture, can be overcome by this method. Running results show a significant increase in accuracy over training models that just use uni-features, with a model having a training accuracy of 99.95% and a validation accuracy of 99.74%. The final result of the model obtained the ability to understand 6 classes of hand gestures made in 3 classes based on the intent of the hand gestures exhibited when participants try the taste of tea, namely Like, Dislike, and Neutral (So-so).

DOI： 10.24507/icicelb.15.08.779

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85202546702&origin=inward

Static Hand Gesture Recognition based on Multi-Feature extraction with Context for User Intension 査読有り国際誌

(1) Khawaritzmi Abdallah Ahmad, Dian Christy Silpani and Kaori Yoshida

ICIC Express Letters, Part B: Applications 15 ( 8 ) 2024年08月

An Investigation of the Effects of Presentation Methods for Visual Elements and Thinking Styles on Text Comprehension 査読有り国際誌

Kouki Ueno, Kaori Yoshida and Naruki Shirahama

ICIC Express Letters, Part B: Applications 15 ( 6 ) 561 - 573 2024年06月

Implicit Attitude of Human Behavior: A Review 査読有り

Dian Christy Silpani and Kaori Yoshida

Proceedings of Computer Information Systems, Biometrics and Kansei Engineering 2023 102 - 114 2023年09月

Static Hand Gesture Recognition Based on Multi-Feature Extraction with Context for User Intension 査読有り国際誌

Khawaritzmi Abdallah Ahmad, Dian Christy Silpani and Kaori Yoshida

Proc. of 17th International Conference on Innovative Computing, Information and Control (ICICIC2023) ( 142 ) 2023年08月

An Investigation of the Effects of Presentation Methods for Visual Elements and Thinking Styles on Text Comprehension 査読有り国際誌

Kouki Ueno, Kaori Yoshida and Naruki Shirahama

Proc. of 17th International Conference on Innovative Computing, Information and Control (ICICIC2023) ( 125 ) 2023年08月

A Feasibility Study on Hand Gesture Intention Interpretation Based on Gesture Detection and Speech Recognition 査読有り国際誌

Silpani Dian Christy, Suematsu Keishi, Yoshida Kaori

Journal of Advanced Computational Intelligence and Intelligent Informatics ( 富士技術出版株式会社 ) 26 ( 3 ) 375 - 381 2022年05月

担当区分：最終著者記述言語：英語掲載種別：研究論文（学術雑誌）

In this paper, considering the characteristics of hand gestures that occur during a natural, unscripted human conversation, we report the results of experiments involving the data analysis of the video recordings of experiments. The recorded conversations between two subjects and the assistant in the experiments (the first author) are analyzed for findings that might be useful to future human-robot interaction (HRI), findings regarding the recognition of hand gestures in natural conversations. In the experiment, hand gestures that appeared naturally during the conversations were manually selected, and their elapsed times and number of appearances were analyzed. In addition, these unscripted conversations were analyzed with a voice recognition system, and the meanings of the hand gestures that appear in the recorded video were interpreted. It was observed that some gestures were made consciously and others unconsciously. Some gestures appeared multiple times with similar meanings and the same intention, especially gestures involving pointing at something to emphasize a sentence or word. Other observations included that there were times when (i) each subject made different hand gestures to convey the same meaning, (ii) both subjects made the same hand gesture but intended different meanings, and (iii) a specific gesture was made by the same subject but with different intended meanings. After considering the results of these experiments, we conclude that it is better to combine gestures with the utterances at the time in order to interpret their intended meanings correctly.

DOI： 10.20965/jaciii.2022.p0375

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85132786254&origin=inward

Hand Gesture Recognition by Hand Landmark Classification 査読有り国際誌

AHMAD Khawaritzmi Abdallah, SILPANI Dian Christy, YOSHIDA Kaori

International Symposium on Affective Science and Engineering ( 日本感性工学会 ) ISASE2022 ( 0 ) 1 - 4 2022年01月

担当区分：最終著者記述言語：英語掲載種別：研究論文（国際会議プロシーディングス）

These years, computer vision technology has been rapidly advanced. On the other hand, customer satisfaction is essential for the industry to improve its service but use traditional methods ineffectively. Using technology such as computer vision, we can now collect the information we are looking for directly from a human. Humans can use many kinds of modalities to interact with computers. Hands are perhaps the largest source of body language information after the face. To understand the gesture's meaning, we can use MediaPipe Hands developed at Google LLC as a method to track and recognize human hands. However, if we want to understand some kinds of hand gestures using MediaPipe Hands, we need to create a condition ourselves by using if-else conditions. This research tried to collect the many varieties of each hand gesture using the 21 key points in x, y, and z coordinates as a feature. As classifiers, we chose the support vector machine (SVM) and the artificial neural network (ANN). This research found that SVM using a polynomial kernel is the best among all of the methods we used as a classifier method for the 3D value of 21 key points from the hand skeleton. The accuracy and F1-score from SVM using a polynomial kernel were 86.26% accuracy and 82% F1-score, respectively, representing the best performance for each class of all the methods we used in this research.

DOI： 10.5057/isase.2022-c000026

その他リンク： https://www.jstage.jst.go.jp/article/isase/ISASE2022/0/ISASE2022_1_34/_pdf

特集：「地域における AI 社会実装」特集「地域における AI 社会実装」にあたって

吉田香, 諏訪博彦, 清田陽司

人工知能 ( 一般社団法人人工知能学会 ) 37 ( 2 ) 170 - 171 2022年01月

Calibration Cost Reduction of Indoor Localization Using Bluetooth Low Energy Beacon 査読有り国際誌

As M., Shimizu H., Benaissa B., Yoshida K., Koppen M.

Journal of Advanced Computational Intelligence and Intelligent Informatics ( 富士技術出版株式会社 ) 26 ( 1 ) 97 - 106 2022年01月

記述言語：英語掲載種別：研究論文（学術雑誌）

Indoor localization based on Bluetooth low energy (BLE) beacons has been rapidly developed, and many approaches have been developed to achieve higher estimation accuracy. In these methods, the received signal strength (RSS) is the input. However, the measurement of indoor environments is affected easily; the signal may be reflected and attenuated by obstacles such as the human body, walls, and furniture, which creates a challenge for methods based on signal mapping. In this study, BLE signal characteristics are investigated in an indoor localization setting. An experiment is performed using one BLE beacon and multiple receivers installed at different wall and ceiling positions. The raw RSS is observed, and the relationship between the BLE beacon signal strength characteristics against the human body effect as well as the receiver's placement in the observation area are discussed. Signal mapping is performed, where the signal strength is measured from all receivers simultaneously. The position estimation accuracy is examined based on different data scenarios. The results show that the estimation position estimated by the BLE beacon based on extensive BLE beacon data does not affect the estimation accuracy.

DOI： 10.20965/jaciii.2022.p0097

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85124460770&origin=inward

A First Prototype of an Emotional Smart Speaker 査読有り国際誌

Enrique A. de la Cal, Alberto Gallucci, José Ramón Villar, Kaori Yoshida, Mario Koeppen

Advances in Intelligent Systems and Computing 1401 304 - 313 2021年09月

A Feasibility Study on Hand Gesture Recognition in Natural Conversation 査読有り

Silpani D.C., Suematsu K., Yoshida K.

2021 5th IEEE International Conference on Cybernetics, CYBCONF 2021 85 - 90 2021年06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）

This paper proposes to approach natural hand gestures that occur in natural conversations. Hand gesture recognition in Human-Robot Interaction is considered a preliminary study for the next step, based on the vision system of a humanoid robot. We experimented as a feasibility study on hand gestures recognition in natural conversation. The conversations between an experimental assistant (the first author) and two different subjects (both female) were recorded in the experiment. Each conversation includes two people, one of whom is the subject of the hand gesture detection that appears during the conversation. We manually captured the hand gestures that appear naturally during the session and analyzed their duration and frequency of occurrence from each video. As a result of conversations with two subjects, it was clear that hand gestures significantly impact people's conversations. We observed that hand gestures with the same word or similar meaning appeared several times with the same intention, such as pointing at something and emphasizing a sentence or word. The topic of natural hand-gesture recognition by a robot will be a challenge. Cultural diversity makes the scope of natural hand-gesture broad, but this can be considered a topic in developing the knowledge of human-robot interaction in the future.

DOI： 10.1109/CYBCONF51991.2021.9464141

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85113788192&origin=inward

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

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

Smart watering system based on framework of low-bandwidth distributed applications (LBDA) in cloud computing 査読有り

Sirimorok N., As M., Yoshida K., Köppen M.

Advances in Intelligent Systems and Computing 1263 AISC 447 - 459 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）

The estimation of soil moisture content is required for agriculture, mainly to build the irrigation scheduling model. In this study, we present a smart watering system to deal with various factors derived from the stochastic information in the agricultural operational, i.e., air temperature, air humidity, soil moisture, soil temperature, and light intensity. The methodology consists of exploiting the Internet of Things (IoT) using Low-Bandwidth Distributed Applications (LBDA) in cloud computing to integrate the real data sets collected by the several sensor technologies. We conducted experiments for the watering system used two types of soil and different plant. Here, the Long Short Term Memory Networks (LSTMs) approach in deep learning techniques used to build smart decisions concerning watering requirements and deal with heterogeneous information coming from agricultural environments. Results show that our models can effectively improve the prediction accuracy for the watering system over various soil and plants.

DOI： 10.1007/978-3-030-57796-4_43

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85090100059&origin=inward

より良い未来のための社会実装

吉田香

人工知能 ( 一般社団法人人工知能学会 ) 36 ( 3 ) 335 - 335 2021年01月

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

プログレスバーの形状が時間評価に及ぼす影響査読有り

大坪正和, 吉田香

バイオメディカル・ファジィ・システム学会誌 ( バイオメディカル・ファジィ・システム学会 ) 18 ( 2 ) 31 - 39 2016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，視覚情報を用いてパーソナルコンピュータのユーザーを満足させることである．パーソナルコンピュータのアプリケーションでは，時間評価に関係する動作が多く存在する．多くの研究は，ユーザーを満足させるために処理速度に焦点を当てている．我々は，ユーザーを満足させるためのアプリケーションインターフェースに着目し，知覚心理学の知見を用いてインターフェースが時間評価に及ぼす影響を調査する．具体的には，グラフィカルインターフェースの一つであるプログレスバーに着目し，プログレスバーの形状が時間評価に及ぼす影響を調査するために主観評価実験を行った．主観評価実験では，異なる10 種類の形状のプログレスバーを実験参加者に提示し，プログレスバーの動作が完了するまでの時間を答えさせた．動作時間は10 秒，12 秒，14 秒の3 種類とした．実験結果より，中心角が90 度のリング型のプログレスバーが時間を一番短く感じさせることが分かった．また，ファイル解凍ソフトを模したアプリケーションを作成し，結果が再現されることを確かめた．今後，より長い時間やより短い時間の処理を待つ場合でも同様の結果が得られるか調査する必要がある．

DOI： 10.24466/jbfsa.18.2_31

その他リンク： https://ci.nii.ac.jp/naid/130007935713

キャラクタの外見特徴と性格印象に応じた音声合成手法の提案

大野歩，大坪正和，吉田香

第 43 回あいまいと感性研究部会ワークショップ 2015年11月

KL 情報量に基づく情報検索性能評価指標の提案

大坪正和，古川徹生，吉田香

第 43 回あいまいと感性研究部会ワークショップ 2015年11月

Towards a Kansei WordNet by Color Design SNS Evaluation 査読有り

Kaori Yoshida, Dwilya Makiwan and Mario Koeppen

Proc. of 2015 International Conference on Intelligent Networking and Collaborative Systems (INCoS2015) 2015年09月

Subjective Speed Evaluation on Striped Image 査読有り

Shunta Ando and Kaori Yoshida

Proc. of the Joint 7th International Conference on Soft Computing and Intelligent Systems and 15th International Symposium on Advanced Intelligent Systems (SCIS&ISIS 2014) 2014年12月

Music Recommendation System based on Lyrics and Melody

Tatsuya Kanatsu, Kaori Yoshida

International Workshop on Machine Vision and Intelligence (IWMVI 2014) 2014年10月

Manga Retrieval System based on Ambiguous Memory

Shogo Nishimura, Kaori Yoshida

International Workshop on Machine Vision and Intelligence (IWMVI 2014) 2014年10月

A Study on Face Impression based on Hair Style

Lu Zhang, Kaori Yoshida

International Workshop on Machine Vision and Intelligence (IWMVI 2014) 2014年10月

Evolving Fair Linear Regression for the Representation of Human-Drawn Regression Lines 査読有り

Mario Köppen, Kaori Yoshida and Kei Ohnishi

Proc. of 2014 International Conference on Intelligent Networking and Collaborative Systems (INCoS 2014) 2014年09月

How does Shape of Progress Bar Effect on Time Evaluation 査読有り

Masakazu Ohtsubo, Kaori Yoshida

Proc. of 2014 International Conference on Intelligent Networking and Collaborative Systems (INCoS 2014) 2014年09月

A Study of Relation between Fixation Point and Subjective Time Evaluation

Yuuki Wakida and Kaori Yoshida

Proc. of the 14th KYUTECH-POSTECH Joint Workshop on Neuroinformatics 2014年08月

A Study of influences on emotional expression by the length of consonant

Tomohito Matsushita and Kaori Yoshida

Proc. of the 14th KYUTECH-POSTECH Joint Workshop on Neuroinformatics 2014年08月

人間と機械が融和するネオファジィの世界

山川烈, 内野英治, 小橋昌司, 田向権, 小出洋, 杉本徹, 椎塚久雄, 吉田香, 吉川大弘, 本田あおい, 林勲, 鈴木倫保, 武川直樹

日本知能情報ファジィ学会ファジィシステムシンポジウム講演論文集 ( 日本知能情報ファジィ学会 ) 30 ( 0 ) 864 - 871 2014年01月

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）

人間同士の，あるいは人間と機械の間のスムーズなコミュニケーションを実現するには，従来のファジィの概念をブレイクスルーした非定量空間で定義される新しい概念，「ネオファジィ」の創出と，その実用技術の開発が必要である．本パネル討論では，①あいまいな言語を使った推論，連想，類推，比喩の意味理解，②テクスト理解，テクスト生成，言語によるコミュニケーションのメカニズム，③ネオファジィ言語を生成・理解する際の脳内ネットワークの形成過程，④ネオファジィ言語と視聴覚表象の対応メカニズム，⑤ネオファジィ専用のメモリやアーキテクチャなど，ハードウェア，⑥ネオファジィ言語を用いたコミュニケーションの応用等について議論する．

DOI： 10.14864/fss.30.0_864

その他リンク： https://ci.nii.ac.jp/naid/130005480616

A Study of Speech Synthesizer with Emotional Expression

Tomohito Matsushita and Kaori Yoshida

Proc. of the International Workshop on Machine Vision for Industrial Innovation (MVII2013) 136 2013年10月

Voice Retrieval based on Subjective Similarity

Masakazu Ohtsubo and Kaori Yoshida

Proc. of the International Workshop on Machine Vision for Industrial Innovation (MVII2013) 137 2013年10月

Illustration Image Retrieval based on Subjective Similarity

Ayumu Ohno and Kaori Yoshida

Proc. of the International Workshop on Machine Vision for Industrial Innovation (MVII2013) 146 2013年10月

Locked Room Mystery Novel Writing Support System

Yuuki Wakida and Nasanobu Umeda and Kaori Yoshida

Proc. of the 13th KYUTECH-POSTECH Joint Workshop on Neuroinformatics 13 - 14 2013年08月

A Study of Speech Synthesizer with Emotional Expression

Tomohito Matsushita and Kaori Yoshida

Proc. of the 13th KYUTECH-POSTECH Joint Workshop on Neuroinformatics 11 - 12 2013年08月

Relational approaches to resource-aware multi-maxmin fairness in multi-valued resource sharing tasks 査読有り

Mario Köppen, Kaori Yoshida and Masato Tsuru

International Journal of Space-Based and Situated Computing 3 ( 2 ) 91 - 101 2013年07月

Meta-heuristic approach to proportional fairness 査読有り

Mario Köppen, Kaori Yoshida, Kei Ohnishi and Masato Tsuru

Evolutionary Intelligence 5 ( 4 ) 231 - 244 2012年12月

Evolutionary Linkage Creation between Information Sources in P2P Networks 査読有り

Kei Ohnishi, Mario Köppen and Kaori Yoshida

Evolutionary Intelligence 5 ( 4 ) 245 - 259 2012年12月

Genetic Clustering Based on Segregation Distortion Caused by Selfish Genes 査読有り

Kei Ohnishi, Mario Köppen, Chang Wook Ahn and Kaori Yoshida

Proc. of 2012 IEEE International Conference on Systems, Man, and Cybernetics 2012年09月

Comparative Study on Meta-heuristics for Achieving Parabolic Fairness in Wireless Channel Allocation 査読有り

Mario Köppen, Kaori Yoshida, Rodrigo Verschae, Masato Tsuru and Yuji Oie

Proc. of 2012 IEEE/IPSJ 12th International Symposium on Applications and the Internet (SAINT) 302 - 307 2012年07月

Evolutionary P2P Networking for Realizing Adaptive Networks 査読有り

Kei Ohnishi, Mario Koeppen, Kaori Yoshida, Yuji Oie

Advances in Intelligent Modelling and Simulation: Artificial Intelligence-based Models and Techniques in Scalable Computing 186 - 213 2012年07月

A Generic Approach to Multi-Fairness and Its Application to Wireless Channel Allocation 査読有り

Mario Köppen, Kaori Yoshida and Masato Tsuru

Proc. Third International Conference on Intelligent Networking and Collaborative Systems (INCoS 2011) 587 - 593 2011年12月

Color Effect on Subjective Perception of Progress Bar Speed 査読有り

Kentaro Hamada, Kaori Yoshida, Kei Ohnishi and Mario Köppen

Proc. Third International Conference on Intelligent Networking and Collaborative Systems (INCoS 2011) 863 - 866 2011年12月

Unsorting the Proportional Fairness Relation 査読有り

Mario Köppen, Kaori Yoshida and Masato Tsuru

Proc. Third International Conference on Intelligent Networking and Collaborative Systems (INCoS 2011) 47 - 52 2011年12月

Recommendation System Based on Competing Algorithms 査読有り

Elnaz Mazandarani, Kaori Yoshida, Mario Köppen and Wladimir Bodrow

Proc. Third International Conference on Intelligent Networking and Collaborative Systems (INCoS 2011) 857 - 862 2011年12月

A GRATIS Theorem for Relational Optimization 査読有り

Mario Köppen, Kaori Yoshida and Kei Ohnishi

Proc. 11th International Conference on Hybrid Intelligent Systems (HIS 2011) 674 - 679 2011年12月

Meta-Heuristic Optimization Reloaded 査読有り

Mario Köppen, Kaori Yoshida and Kei Ohnishi

Proc. Third World Congress on Nature and Biologically Inspired Computing (NaBIC2011) 562 - 568 2011年10月

Fuzzy Approaches to Proportional Fairness 査読有り

Mario Köppen, Kaori Yoshida and Masato Tsuru

2011 International Conference on Emerging Intelligent Data and Web Technologies (EIDWT) 72 - 77 2011年09月

A Heuristic Approach to Fair Routing Path Selection 査読有り

Noriaki Hirosue, Kaori Yoshida and Mario Köppen

15th International Conference of Knowledge-Based and Intelligent Information and Engineering Systems (KES2011) 176 - 186 2011年09月

Annealing Heuristic for Fair Wireless Channel Allocation by Exponential Ordered-Ordered Weighted Averaging Operator Maximization 査読有り

Mario Köppen, Kaori Yoshida, Masato Tsuru and Yuji Oie

2011 IEEE/IPSJ International Symposium on Applications and the Internet (SAINT 2011) 538 - 543 2011年07月

Interactive genetic algorithm for font generation system 査読有り

Yoshida K., Nakagawa Y., Köppen M.

2010 World Automation Congress, WAC 2010 2010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）

We suggest the font generation system consists of font definition phase and font creation phase. For font definition phase, the tool to export font element parameters represented by XML format has implemented. For font creation phase, the font creation application PALETTE (Personal Adapted LETTEr) has designed to emerge various fonts based on user's Kansei without hand drawing and developed as Adobe AIR application. It enables to generate a new font by selecting user's favorite fonts from displayed candidates, save font as an image according to users input text, and reuse stored font to the next font creation. To design PALETTE, we focused on Interactive Genetic Algorithm (IGA) technique to take user's subjective desire on font shape into the system. IGA is well known as a method to use user's subjective evaluation interactively. PALETTE has the following modules; (i) create new font candidates base on general font, or stored font characteristics with Kansei label, (ii) generate new fonts using genetic algorithm which can modify each element of the selected candidates, and (iii) export font image and save font characteristics. In this paper, we explain the font generation system and show the experimental results. © 2010 TSI Press.

その他リンク： https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=78651447534&origin=inward

Fuzzification of maxmin fairness relation based on subvector dominance degree 査読有り

Köppen M., Yoshida K., Tsuru M.

Proceedings - 2nd International Conference on Intelligent Networking and Collaborative Systems, INCOS 2010 344 - 349 2010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）

Maxmin fairness can be characterized as a state obtained from the application of the bottleneck flow control algorithm, allowing for a formally efficient definition of a fairness concept. But a number of problems with the related maxmin fairness vector relation, esp. non-uniqueness of the maximum, preference for improving least components, artefacts appearing in the discrete domain, and lack of numerical representation have already been noted. For this reason, we propose a fuzzy fairness, based on subvector dominance. The fuzzy fairness, here called subvector maxmin dominance degree, is based on the ratio of the number of all maxmin fair dominating subvectors to the number of all possible subvectors. It is shown how this definition extends the maxmin fairness relation, how it helps to solve the problems with maxmin fairness, and how it numerically emphasises fairness states against other states. © 2010 IEEE.

DOI： 10.1109/INCOS.2010.45