Talking Face Video Super-Resolution

MPhil Thesis Defence


Title: "Talking Face Video Super-Resolution"

By

Mr. Chang Dae PARK


Abstract

This work explore the reference-based talking face video super-resolution 
for video conferencing even under low bandwidth network condition. Our 
objective is to reconstruct high quality talking face video with given low 
resolution video and sparsely given high resolution frames for every 10 
frames. To this end, our method utilize the pretrained GANs as a prior 
knowledge to reconstruct photo-realistic face images. Using GANs 
pretrained on large dataset is much helpful to generate plausible face 
images even with the low resolution images, however, it show low fidelity. 
It means that the person’s face identity between original and 
reconstructed ones are quite different. To tackle this problem, our method 
is designed to exploit the multiple high resolution feature which can help 
generate high fidelity face images. The proposed method exploits the 
recent development of reference-based super-resolution techniques and we 
modify to enable our model to utilize more than single reference image. 
Experimental results show that the proposed method can generate high 
fidelity talking face video when more reference frames are given.


Date:  			Monday, 23 August 2021

Time:			2:00pm - 4:00pm

Zoom meeting:
https://hkust.zoom.us/j/93042566103?pwd=VXRFY3pFaHo0bm5WMnRYeWhCWkVSdz09

Committee Members:	Dr. Qifeng Chen (Supervisor)
 			Dr. Hao Chen (Chairperson)
 			Dr. Dan Xu


**** ALL are Welcome ****