Downloads: 236

Files in This Item:
File Description SizeFormat 
yjohk00730.pdfAbstract_要旨172.61 kBAdobe PDFView/Open
gjohk00730.pdfDigest_要約23.01 kBAdobe PDFView/Open
Title: Deep Learning Approaches on the Recognition of Affective Properties of Images
Other Titles: 深層学習を用いた画像の情動的属性の認識
Authors: Yamamoto, Takahisa
Author's alias: 山本, 卓永
Keywords: Deep Learning
Image Emotion Recognition
Fashion Style Recognition
Semantic Features
Fine-Grained Features
Issue Date: 23-Sep-2020
Publisher: Kyoto University
Conferring University: 京都大学
Degree Level: 新制・課程博士
Degree Discipline: 博士(情報学)
Degree Report no.: 甲第22800号
Degree no.: 情博第730号
Conferral date: 2020-09-23
Degree Call no.: 新制||情||125(附属図書館)
Degree Affiliation: 京都大学大学院情報学研究科知能情報学専攻
Examination Committee members: (主査)准教授 中澤 篤志, 教授 西野 恒, 教授 鹿島 久嗣
Provisions of the Ruling of Degree: 学位規則第4条第1項該当
Rights: 学位規則第9条第2項により要約公開
Chapter 4 (from Section 4.1 to 4.4.1) (c)IEICE. Chapter 4 (from Section 4.1 to 4.4.1) is based on the work published in Takahisa Yamamoto and Atsushi Nakazawa: Fashion Style Classification through Separating an Input Image into Elements and Elements-based Learning, The Transaction of The Institute of Electronics, Information and Communication Engineers, vol. J102-D, no. 10, pp. 639-650, October 2019 (in Japanese). DOI:10.14923/transinfj.2019IEP0004; Chapter 4 (Section 4.5) (c)IEEE. Chapter 4 (Section 4.5) is based on the work published in Takahisa Yamamoto and Atsushi Nakazawa: Fashion Style Recognition Using Component-Dependent Convolutional Neural Networks, IEEE International Conference on Image Processing (ICIP), pp. 3397-3401, September 2019. DOI: 10.1109/ICIP.2019.8803622;
DOI: 10.14989/doctor.k22800
URI: http://hdl.handle.net/2433/259068
dc.relation.haspart: https://doi.org/10.14923/transinfj.2019IEP0004
https://doi.org/10.1109/ICIP.2019.8803622
Appears in Collections:140 Doctoral Dissertation (Informatics)

Show full item record

Export to RefWorks


Export Format: 


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.