Oriel Frigo

Bio

Hi, welcome!

I am currently a Technical Lead Researcher at Huawei Von Neumann Research Center, Zurich, Switzerland. I am working towards efficient and Ascend-native multimodal generation.

Between 2023 and 2026 I worked as Lead Machine Learning Researcher at Brahma AI (previously Metaphysic), where I applied research on generative models (VAEs, GANs, DiTs) to produce visual effects for Hollywood movies such as "Here", "Alien: Romulus" and "Mad Max: Furiosa". I was delighted to receive the 23rd VES award for Emerging Technology for the movie "Here" as recognition for this applied research work.

Between 2019 and 2022, I have worked on R&D in the field of AI / machine learning at AnotherBrain (Paris). There I have conducted research on VAEs for anomaly localization and led a research group for Heliaus, an European funded project on thermal-color semantic segmentation for autonomous driving applications.

In 2018 I was working with neuromorphic / event-based computer vision R&D at Prophesee (Paris). In 2017 I did a postdoc at INRIA Rennes, working with Christine Guillemot on Light field editing. Between 2013 and 2016 I was a PhD student, working on “Example guided video editing”, within a thesis agreement between Paris Descartes University (supervised by Julie Delon) and Technicolor (supervised by Pierre Hellier and Neus Sabater).

In July 2013, I completed the Master program "Color in Informatics and Media Technology" (CIMET). During this period, I have studied at Jean Monnet University (Saint-Etienne, France) and University of Eastern Finland (Joensuu, Finland). My master thesis internship took place at Technicolor, on the topic of “Example guided color transfer”.

In 2010 I received the BSc degree in Computer Science from the University of the State of Santa Catarina (Brazil), with a dissertation entitled " Automatic Classification of Engravings Based on Texture Analysis and Self-organizing Maps ".

My research interests include Machine Learning, Generative Models, VAEs, DiTs.

In my free time, I like to play guitar, to swim, and to skate.

Contact: oriel dot frigo at gmail dot com

Publications

Journal papers:

  • O. Frigo, J. Delon, N. Sabater, P. Hellier, Video Style Transfer by Consistent Adaptive Patch Sampling. The Visual Computer, 2018. [pdf] [site]
  • O. Frigo, J. Delon, N. Sabater, P. Hellier, Motion Driven Tonal Stabilization, IEEE Transactions on Image Processing, 2016. [pdf] [site]

Conference papers:

  • J Plaete, M Olivieri-Dancey, O Frigo, M Anton, S Correa, S Deckers, Neural Performance Toolset: AI-Powered Human Performance Synthesis. SIGGRAPH Talks, 2025
  • O.Frigo, L. Martin-Gaffe, C. Wacongne DooDLeNet: Double DeepLab Enhanced Feature Fusion for Thermal-color Semantic Segmentation, CVPR Workshops, 2022. [pdf]
  • D. Dehaene*, O. Frigo*, S. Combrexelle, P. Eline (* = equal contrib.) Iterative energy-based projection on a normal data manifold for anomaly localization, ICLR, 2020. [pdf] [site (OpenReview)]
  • O. Frigo, C. Guillemot, Epipolar Plane Diffusion: An efficient approach for Light Field Editing, BMVC, 2017. [pdf] [site]
  • O. Frigo, J. Delon, N. Sabater, P. Hellier, Split and Match: Example-based Adaptive Patch Sampling for Unsupervised Style Transfer, CVPR, 2016. [pdf] [site]
  • O. Frigo, N. Sabater, J. Delon, P. Hellier, Motion Driven Tonal Stabilization, ICIP, 2015. [pdf] [site]
  • O. Frigo, N. Sabater, J. Delon, P. Hellier, Stabilization tonale de video, GRETSI, 2015. [pdf] [site]
  • O. Frigo, N. Sabater, V. Demoulin, P. Hellier, Optimal Transportation for Example-Guided Color Transfer, ACCV, 2014. [pdf] [site]
  • P.-A. Bokaris, O. Frigo, A. Chen, I. Ciortan, A. Pelegrina, Basic and Advanced Colorimetry Methods for Displaying Microscope Image Appearance, In Proc. 12th International AIC Congress, 2013.
  • O. Frigo, A. G. Silva, A. Nied . Classificação Automática de Gravuras Baseada em Análise de Texturas e Mapas Auto-Organizáveis. III Workshop on Computational Intelligence (WCI), 2010. p. 446-451. (in portuguese)

Preprints:

  • R. Brossard, O. Frigo, D. Dehaene, Graph convolutions that can finally model local structure. ArXiv, 2021. [pdf] [code]
  • O. Frigo, R. Brossard, D. Dehaene, Graph Context Encoder: Graph Feature Inpainting for Graph Generation and Self-supervised Pretraining. ArXiv, 2021. [pdf]
  • R. Brossard, O. Frigo, D. Dehaene, Realistic molecule optimization on a learned graph manifold . ArXiv, 2021. [pdf]

PhD Thesis:

  • O. Frigo, Example-guided Video Editing. Paris Descartes University, 2016. [pdf]