콘텐츠 소개(About Contents)

Lip Reading is an important technique for many people as well
as various situations that involve communication. With the ad- vance in computer vision and deep neural networks, lip reading has started to become automated. This project has developed a working model based on 3D CNN and GRU for lip reading of spo- ken words, without auditory information. Although there is room for improvement, this model can serve as a comparative standard for future lip reading models.
KEYWORDS
visual  speech  recognition,  lip  reading,  CV,  3D  CNN,  LSTM,  GRU

기획자(Planner)

서울대학교 자유전공학부      이서영