A Survey on Video-based Sign Language Recognition and Translation

PhD Qualifying Examination


Title: "A Survey on Video-based Sign Language Recognition and Translation"

by

Mr. Ronglai ZUO


Abstract:

Sign languages are the principal communication method among 
hearing-impaired people. However, few normal hearing people can master 
them, and thus there is a communication gap between the deaf and the 
normal hearing people. Sign language recognition and translation (SLR and 
SLT) aim to fill in this gap by automatically transcribing or translating 
a sign language video into glosses (sign language words) or spoken 
language words. The two tasks are so challenging that both the knowledge 
of computer vision and natural language processing are needed. Also, the 
video data suffer from occlusion and motion blur problems, which make both 
tasks more difficult. In recent years, deep-learning-based techniques have 
significantly boosted the performance of SLR and SLT models, although it 
is still not good enough for real practice. In this survey, we will 
elaborate on the framework adopted by most SLR and SLT works, which 
consists of a feature extractor and a sequence model, along with the 
limitations of existing works. Finally, we discuss some possible research 
directions.


Date:			Monday, 20 December 2021

Time:                  	2:30pm - 4:30pm

Venue:			Room 4472
 			(lifts 25/26)

Committee Members:	Dr. Brian Mak (Supervisor)
 			Prof. Raymond Wong (Chairperson)
 			Dr. Dan Xu
 			Prof. Nevin Zhang


**** ALL are Welcome ****