We are looking for ambitious and motivated candidates for 18 months post-doctoral position in digital avatars, computer vision and machine learning.
Digital avatars can engage with humans to interact socially. However, to design a realistic digital avatar with a natural interaction, the most important challenge is to understand how the changes in facial expressions and postures influence the social interactions. We will combine methods from computer vision, machine learning and cognitive sciences to design “an empathic” digital avatar which looks like human beings and can engage a natural interaction with them. This digital avatar should have spontaneous dynamic movements of the face and postures.
In this project we will develop a data-driven approach. Indeed, over the past few years major developments in deep learning have enabled important advancements in artificial intelligence. As in many other computer vision tasks, deep learning has brought revolutionary advances in human behaviour understanding from visual data. Deep models are now effective not only in detecting and recognizing human faces, actions and activities but also in generating realistic human-like behavioral data. This project will attempt to achieve the following scientific and technological goals:
The position is expected to be filled before 31th July 2020. Please send the following information in a single PDF file to Professor Mohamed Daoudi http://www.cristal.univ-lille.fr/~daoudi, firstname.lastname@example.org subject [PostDocAvatar]:
1.Xavier Alameda-Pineda, Elisa Ricci, Albert Ali Salah and Nicu Sebe and Shuicheng Yan, Special Issue on Generating Realistic Visual Data of Human Behavior, Int. J. Comput. Vis., Vol. 128, No. 5, pp. 1376--1377, 2020.
2.Caroline Chan, Shiry Ginosar, Tinghui Zhou, Alexei A. Efros: Everybody Dance Now. ICCV 2019: 5932-5941
3.Naïma Otberdout, Mohamed Daoudi, Anis Kacem, Lahoucine Ballihi and Stefano, Dynamic Facial Expression Generation on Hilbert Hypersphere with Conditional Wasserstein Generative Adversarial Nets," in IEEE Transactions on Pattern Analysis and Machine Intelligence, doi: 10.1109/TPAMI.2020.3002500. (2020) https://arxiv.org/pdf/1907.10087.pdf
4.Eli Shlizerman, Lucio M. Dery, Hayden Schoen, Ira Kemelmacher-Shlizerman: Audio to Body Dynamics. CVPR 2018: 7574-7583.
(c) GdR 720 ISIS - CNRS - 2011-2020.