Annonce

Les commentaires sont clos.

Synthesis of videos driven by text and audio

17 Janvier 2022


Catégorie : Post-doctorant


Key words: Video synthesis, generative adversarial networks, text to speech to video methods, image and video analysis and processing

 

During this post-doctoral contract, we want to work on the development of a prototype of a Text to Speech to Video technology with a sufficient level of accuracy.

The final goal is the construction of a Text to Speech to Video technology allowing to generate in a totally automatic way, several minutes long video sequences of a person speaking in front of a camera (talking head) from a textual script. A generated video must have a sufficient quality to allow the illusion of an original video.

In the first step, the research will focus on the generation of photorealistic videos of a person's face and mouth. In the second step, the development of the Text to Speech to Video solution will be proposed to allow the word change, word deletion, and the word addition not necessarily pronounced by the agent. And in the third step, we will work on improvement of photorealism of the whole video.

The funding covers 18 months of post-doc, the desired start is April 2022.

Contact e-mails:

iuliia.tkachenko@univ-lyon2.fr

thibault@monpetitplacement.fr

More details https://perso.liris.cnrs.fr/itkachenko/public_html/2022_post_doc_eng.pdf