Slim Ouni · LORIA · Inria

Université de Lorraine · LORIA · Inria

Slim Ouni

Professor of Computer Science. Head of the MULTISPEECH research team — a joint team of Inria, CNRS and Université de Lorraine.

I work on multimodal speech communication: how the acoustic signal, the movements of the articulators, facial expression and co-speech gesture are produced together, and how machines can analyse and generate them jointly. That question runs from articulatory modelling and expressive audiovisual speech synthesis to co-speech gesture generation, sign language production, and clinical applications of multimodal speech analysis.

Université de Lorraine LORIA Inria CNRS
Portrait of Slim Ouni

What our work looks like

Research →
Automatic co-speech gesture generation Full-body gesture generated from speech alone, then retargeted onto a virtual character. PhD work of Louis Abel (2021–2025)·
MetaHuman animation with DeepLipsync Speech-driven facial animation on Epic Games MetaHuman characters, produced during the EPIC MegaGrants project. EPIC Games MegaGrants, 2022–2023·
Realistic video dubbing Re-synchronising the lip movements of a filmed speaker with a soundtrack in a target language. Collaboration between Interdigital & Inria within EU AI4EU (2019-2021)
Automatic lipsync of 3D characters Facial animation generated automatically from audio and text, with no manual keyframing. Technology transferred to Dynalips (2023)·

Research in brief

All five directions →

Direction 01

Expressive audiovisual speech

Synthesising and analysing speech together with the face that produces it — coarticulation, expressivity, intelligibility of talking heads.

Direction 02

Speech and gesture in interaction

Modelling how gesture and speech are synchronised, and generating co-speech gesture from the speech signal.

Direction 03

Sign language processing

End-to-end speech-to-sign generation and the representations that make it learnable — pose, phonology, prosody.

News

  • The Inria–DFKI – ROGSILT (Robust and Generalizable Sign Language Translation) - started (April 2026). I'm Co-PI. A France–Germany programme combining NMT, speech processing and computer vision.
  • The ANR project Syncogest (2025–2029), on speech and gesture synchronisation, has started. I coordinate it.

Open positions

I regularly welcome inquiries from prospective PhD students and postdoctoral researchers. If no position is currently advertised, feel free to contact me directly with your CV and a short statement of research interests. Depending on your profile and interests, we can explore possible research topics and funding opportunities.