ホウチン　テルヒサ   Teruhisa Hochin   寶珍　輝尚    所属   追手門学院大学  理工学部 情報工学科    職種   教授
発行・発表の年月	2021/06/20
形態種別	論文
査読	査読あり
標題	Facial Expression Intensity Estimation using Deep Convolutional Neural Network.
執筆形態	共著・編著（代表編著を除く）
掲載誌名	ACIT 2021: The 8th International Virtual Conference on Applied Computing & Information Technology(ACIT)
掲載区分	国外
出版社・発行元	ACM
巻・号・頁	7-12頁
担当区分	最終著者
著者・共著者	Tomoki Kawashima,Hiroki Nomiya,Teruhisa Hochin
概要	Facial expression recognition has been applied in various fields, but it is not enough to express complex emotions. It is required to detect the intensity of facial expression because the intensity of facial expression varies depending on the intensity of the emotion. In this research, we propose a method for estimating the facial expression intensity using deep convolutional neural networks (deep CNN). First, we collect facial expression intensity labels by experiments and evaluate the accuracy of the labels. Next, we train the deep CNN model by fine-tuning due to the small dataset. As a result, we show that the deep CNN method is effective compared to the results of hand-crafted facial features. However, there are still issues of estimation accuracy for facial expressions with small data and overfitting due to shortage of data.
DOI	10.1145/3468081.3471060
DBLP ID	conf/acit2/KawashimaNH21
PermalinkURL	https://dblp.uni-trier.de/rec/conf/acit2/2021
researchmap用URL	https://dblp.uni-trier.de/db/conf/acit2/acit2021.html#KawashimaNH21