Faculty Profiles - HARADA Toshinobu

Published Papers - HARADA Toshinobu

Division display1 - 20 of about 87／ All the affair displays >>

TF-IDF 併記決定ルール群表示と形式概念分析を用いた決定ルール群構造可視化システムの開発

野口七海, 中村鴻成, 原田利宜（Part： Corresponding author )

日本感性工学会論文誌 2025.03 [Refereed]

Feature Extraction using Deep Learning and Analyses of Curvature on Facial Shapes across Two Races and between Males and Females

Daiki YAMADA, Toshinobu HARADA, Akira YAMADA, Nikhil SHAH, Emily CHWA, Sophia ALLISON （Part： Corresponding author )

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 2024.02 [Refereed]

DOI

Comparative Study between the Exploration of Diverse Solutions by CVAE and JMVAE for Eyeglass Design

Katsuki MOROMUGI, Toshinobu HARADA （Part： Corresponding author )

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 22 ( 1 ) 1 - 10 2023.01 [Refereed]

DOI

Feature Analysis of Impressions around Women’s Eyes and Eyebrows using Neural Network

Yusuke UENO, Toshinobu HARADA, Yohei ISHIGURO （Part： Corresponding author )

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 22 ( 1 ) 11 - 20 2023.01 [Refereed]

DOI

ラフ集合と形式概念分析を用いた包含関係併記決定ルール群表記法の提案

中村鴻成, 原田利宣（Part： Last author )

日本感性工学会論文誌 21 ( 1 ) 113 2022.03 [Refereed]

深層生成モデルによる印象語からのメガネデザイン生成システムの開発

諸麦克紀, 原田利宣（Part： Last author )

日本感性工学会論文誌 21 ( 1 ) 23 - 32 2022.03 [Refereed]

DOI

Analyses of characteristics of Japanese forehead shapes for Masculine / Feminine using deep learning

Saki OTSUDA, Toshinobu HARADA （Part： Last author )

日本感性工学会論文誌 2021.03 [Refereed]

Development of Simultaneous Inclusion and Correlation Visualization System Using Formal Concept Analysis and Quantification Theory Class III

SHIMADA Yusei, HARADA Toshinobu

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 2019 [Refereed]

<p>In recent years, while data mining has attracted attention, formal concept analysis has the disadvantage that although it can understand the inclusion relation between attributes, it doesn't know what kind of meaning or inclusion relation the attributes have. Therefore, in this study, the inclusion of the attributes and the visualization of the correlation are realized by combining the Hasse diagram by formal concept analysis and the layout diagram of the attribute of quantification theory type III for visualizing the meaning between the attributes. Moreover, in order to confirm usefulness of the proposed method, we developed a simultaneous visualization system of inclusion and correlation combining formal concept analysis and quantification theory type III, and evaluated the proposed method using verification data. As a result, it was confirmed that useful information that can't be obtained only by formal concept analysis or quantification theory type III can be extracted.</p>

DOI

Analyses and Visualization of Characteristics of Car Body Types by Using Convolutional Neural Network

TANAKA Syuntaro, HARADA Toshinobu, ONO Kenji

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 18 ( 1 ) 113 - 121 2019 [Refereed]

The cars are classified by cars' body types. However the characteristics are basically similar at first sight, so it is difficult to distinguish the differences among those cars' body types. Therefore, in this study, we considered that cars' characteristics could be analyzed by using deep learning and image recognition technology, developed a system to visualize the judgment and characteristic parts of cars' body types. Specifically, we made renderings of the CG model of 30 cars by setting 360 viewpoints in 1 degree increments around each car. Deep learning was performed using these 2D images as teacher signals. The car body type recognition probability of each angle is graphed, and the characteristic parts of each car body type are visualized. As a result, we clarified the visual angles and the pars contributing the judgment of cars' body types.

DOI

Preference Analyses of Luxury Brands' Web Page Designs using Conceptual Structure Visualization System

SAKOU Issei, SHIRAKO Chie, HARADA Toshinobu

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 17 ( 4 ) 425 - 435 2018 [Refereed]

In previous studies, we developed the conceptual structure visualization system (CSVS) by combining rough set theory and formal concept analysis. Furthermore, using CSVS, we found “conflict effect”, “neutralizing effect” and “separation effect” as characteristic conceptual structures that were difficult to find by using conventional methods. In this study, we aimed to verify the practicality of CSVS by analyses for data that is larger than data on previous studies, and to find the characteristic conceptual structure by using CSVS. First, we selected web page designs of luxury brands as case studies, and we conducted questionnaire survey related to preference for 12 web page designs with 46 female subjects. Next, we visualized and analyzed characteristic of conceptual structures of subjects' preferences using CSVS. And we confirmed that CSVS was available to find conceptual structures of preferences. In addition, we discovered new characteristic conceptual structure like compound of “conflict effect” and “neutralizing effect”.

DOI

Clarification of the Relationship between Proportion Items and Impressions of 1.5 BOX Mini-car Design

MINAMIMOTO Sho, HARADA Toshinobu

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 17 ( 1 ) 139 - 148 2017 [Refereed]

Designers consider cars' proportions using their past experiences and knowledges. However, designers may not search for all patterns of proportion satisfy requirements. In this paper, we aimed to clarify relationships between mini-cars' proportion items and users' impressions. First, we clarified correlations among the proportion items of existing 25 mini-cars using principal component analysis, and created 9 mini-cars' images using the design of experiments. Second, we conducted impression evaluation experiments for 95 subjects using the images. As a result, we clarified relationships between the proportion items and subjects' impressions. Third, we classified subjects into 6 clusters using the difference of their preferences, and clarified the patterns of proportion items that each cluster's subjects feel attractive. Finally, we conducted the evaluation re-experiment for 2 clusters' 20 subjects using the images with the patterns of proportion items preferred by them. As a result, we found strong correlations between mini-cars' heights and their preferences.

DOI

Comparison and Visualization of Affective Information on Two Liquors' Reviews on the Web Using Data Mining

TAKENOSHITA Ryo, HARADA Toshinobu, ONO Kenji

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 17 ( 1 ) 11 - 20 2017 [Refereed]

In the plan of the products, affective information is utilized with diversification of a life style. Then, in this research, we developed the system that visualizes the difference in affective layered structures of two liquors' reviews (hereafter, visualization system). Specifically, the development consists of the following steps. (1) The visualization system extracts affective words from the reviews on the web. (2) We create the decision table in rough set theory using affective words. (3) The visualization system visualizes the difference in affective layered structures of two liquors using decision rules from each decision table. We conducted the evaluation experiment of the visualization system for 10 subjects. We made five subjects summarize comparison results of two liquors' reviews using the visualization system, and made remaining five subjects summarize them without it. Consequently, the usefulness of the visualization system was verified from the summarized contents and the creation time of them.

DOI

Development of Conceptual Structure Visualization System Using Rough Set Theroy and Formal Concept Analysis

SAKOU Issei, HARADA Toshinobu （Part： Last author )

Bulletin of Japanese Society for the Science of Design ( Japanese Society for the Science of Design ) 64 ( 3 ) 3_11 - 3_20 2017 [Refereed]

Rough set theory (RST) can get decision rules (DRs) that express relationship between condition attributes and decision attributes. However, RST has a weak point that it is difficult to grasp conceptual structures (inclusion relations) among several DRs. Therefore, the purpose of this study was to combine RST and formal concept analysis (FCA) in order to make up for the weak points of RST, and propose analysis to find causal relation and inclusion relation that it is difficult to find by using two methods individually. Next, we developed a conceptual structure visualization system that executes two analytical methods simultaneously and projects DRs on Hasse diagram. Lastly, we verified usefulness of the system through comparing the results of analyses by using the system with the results without the system by using data in papers dealing with FCA or RST. And we confirmed that we can easily find conceptual structures not to be found by conventional methods with the system.

DOI

Visualization of Affective Layered Structures of Liquors' Reviews on the Web using Rough Set Theory

TAKENOSHITA Ryo, HARADA Toshinobu, ONO Kenji

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 16 ( 1 ) 19 - 28 2017 [Refereed]

In the plan of products, affective information is increasingly utilized with diversification of a life style. Then, in this research, we had developed the system that visualizes the affective layered structures of liquors' reviews using rough set theory (hereafter, visualization system). Specifically, the development consists of the following steps. (1) The visualization system extracts affective words from the reviews of mail order sites. (2) We create the decision table in rough set theory using affective words. (3) We visualize the affective layered structure using the computed decision rules from decision table. We conducted the evaluation experiment of the visualization system for 20 subjects. We made ten subjects the summary of liquor reviews using the visualization system, and made the remaining ten subjects the summary of them without the visualization system. Consequently, the usefulness of the visualization system was verified from the contents and the time for creation of the summary.

DOI

Study on the Attraction and the Comprehensibility of Icons Using Rough Set

Oono Shintaro, Kaneda Yukihiro, Harada Toshinobu

Bulletin of Japanese Society for the Science of Design ( Japanese Society for the Science of Design ) 62 ( 6 ) 6_61 - 6_68 2016 [Refereed]

By the spread of tablet in recent years, we can see a large variety of icons. However, the comprehensibility and the attraction of icons are different by the cognitive process and preference of the user. Therefore, we aimed to analyse the attraction and the comprehensibility of icons using the rough set in this study. Specifically, at first we investigated the attraction and comprehensibility of icons and extracted form elements (attribute values) to constitute icons. Next, we calculated decision rules of each subject using a rough set to clarify how the attribute values that we extracted influences the attraction and the comprehensibility. In addition, we classified subjects using cluster analysis by using co-occurrence rate between subjects with decision rules. Next, we extracted characteristic attribute values that subjects in each cluster felt attracted and comprehended. Finally we produced sample icons by using the characteristic attribute values and verified availability of these attribute values.

DOI

Analyses of the Characteristics of Japanese Bilateral Upper Eyelid Curves and Impressions of Masculine/Feminine

WATANABE Ryusuke, KAWAKAMI Mari, HARADA Toshinobu

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 15 ( 3 ) 377 - 386 2016 [Refereed]

The eyes play a crucial role in defining the person's impression. However, there are still few studies of evaluation for the impressions of eye's form including the curve of bilateral upper eyelid. Thus, the aims of our study were to analyze and systematize the curve of bilateral upper eyelid using images of 100 Japanese subjects' eyes. Consequently, we classified the curves of males into 5 clusters and females into 6 clusters. Additionally, our study clarified the relationship between the impressions of masculine/feminine and the eye's form elements that consisted of the curves systematized the ratio of the width to height, etc. Concretely, we made 70 subjects to evaluate samples assembled with the eye's form elements by using design experiment. As a result of analysis of variance and multiple comparisons of main effects, we clarified that the bilateral upper eyelid were important form elements affecting the person's impression of masculine/feminine.

DOI

Analyses of the Comprehensibility and the Impressions of Dynamic Pictograms Using Color Expressions

Shintaro Ono and Toshinobu Harada （Part： Last author )

IASDR2015 2015.11 [Refereed]

Analyses of the characteristics of Japanese facial curves and creating the templates to improve outcomes in plastic surgical reconstruction

Toshinobu Harada, Akira Yamada

Journal of the Institute of Image Electronics Engineers of Japan ( Institute of Image Electronics Engineers of Japan ) 44 ( 3 ) 533 - 538 2015

Analyses of the Characteristics of Forehead Curves in Japanese Women and Creating Curve Templates to Improve Outcomes in Plastic Surgical Reconstruction

WATANABE Ryusuke, TAKEUCHI Mai, HARADA Toshinobu, YAMADA Akira

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 14 ( 1 ) 197 - 206 2015 [Refereed]

Creating beautiful curves of the forehead is considered to be the key to achieve optimal outcome of craniofacial plastic surgical reconstruction. However only subjective, or artistic techniques have been proposed in the past, therefore the outcomes of reconstruction seem to be entirely dependent on surgeon's ‘kansei’ or skills. Thus the aim our study were to analyze the curves of the forehead in 50 normal Japanese women, and classify the curves to create as a guide for total forehead reconstruction. To characterize the individual nature of the complex forehead curves, two key lines were abstracted from 3D data. Based on profile of curvature analysis, forehead curves were classified into 6 types in horizontal section line, and 2 types in meridian line. These classification were then used to develop curve templates as a guide for forehead construction. We applied these curve templates in complex craniofacial reconstruction and found that the outcome is very promising.

DOI

Abstraction of Visual Languages in Pictograms Using Color Expression and its Verification

ONO Shintaro, UENISHI Ayaka, HARADA Toshinobu

Transactions of Japan Society of Kansei Engineering ( Japan Society of Kansei Engineering ) 14 ( 3 ) 391 - 400 2015 [Refereed]

Pictograms attract attention in supporting the non-verbal communication and using of interface icon. The existing pictograms are basically painted in monochrome. It was clarified from existing studies that we could clear the mean of pictograms using color expression. However, there were few studies which analyzed the relation between comprehensibility and color express of pictograms and few useful design indexes when painting colors of the pictogram. Then, this study aims to analyze comprehensibility of the pictograms using color expression and to abstract the visual languages. The visual languages are visual elements that composing the design of pictograms communicating information visually. Concretely, we clarified typical colors and color combinations of morphological elements in 20 pictograms conveying meanings of verbs. On the basis of this result, we investigated the relation between comprehensibility and color expression of pictograms by experimental design method. And, we abstracted the visual languages for painting colors of the pictogram and verified its usefulness. As a result, we have clarified the usefulness of some visual languages. However, the comprehensibility of some pictograms was decreased by colorization, we considered characteristics of those color expression.

DOI

　PREV- NEXT 　

<HARADA Toshinobu>

Published Papers - HARADA Toshinobu