KSC 2020
Current Result Document :
| 한글제목(Korean Title) |
Automatic Facial Expression Generation for Sign Language with Neural Machine Translation |
| 영문제목(English Title) |
Automatic Facial Expression Generation for Sign Language with Neural Machine Translation |
| 저자(Author) |
Eui Jun Hwang
Jung-Ho Kim
Jong C. Park
|
| 원문수록처(Citation) |
VOL 47 NO. 02 PP. 0322 ~ 0324 (2020. 12) |
한글내용
(Korean Abstract) |
|
영문내용
(English Abstract) |
In a sign language, facial expressions play an important role for effective communication. In particular, they are well known for conveying emotional and grammatical information that affects the meaning of a sign. In this paper, we only consider the grammatical functions of the facial expressions. We propose a transformer-based facial expression generation model that translates an expression in spoken language into continuous facial landmark sequences for sign language. In order to train the model efficiently, we employ Principal Component Analysis embedding and Custom Mean Squared Error loss. We report the quantitative and qualitative results of the generated facial landmarks.
|
| 키워드(Keyword) |
|
| 파일첨부 |
PDF 다운로드
|
