Maoyuan Xu

prof_pic_me.png

Chengdu, Sichuan, China.

As an R&D Scientist at Ubisoft La Forge, I specialize in Machine Learning, Computer Vision, and Computer Graphics, with a strong focus on advancing the frontiers of game asset creation. My work spans low-level image processing, high-level vision tasks, and the application of diffusion models, all aimed at developing innovative technologies for asset generation and processing.

I expertise in AI-Generated Content (AIGC) and High-fidelity Asset Processing to achieve production-ready quality contents. My interests lie in applied research: identifying real-world production challenges, designing novel solutions, and deploying these innovations into practical tools and workflows to enable more refined and controllable content generation.

In addition to my research, I have the privilege of mentoring research students during their internships, providing hands-on guidance in those rapidly advancing fields. We are consistently looking to forge collaborations with motivated Master’s and PhD candidates who are passionate about pushing the boundaries of knowledge and contributing to novel solutions in Generative AI and Computer Vision.

news

Aug 20, 2025 Our new paper for PBR material super-resolution has been submitted to AAAI-2026.
Jul 08, 2021 Our paper, entitled NFCNN: Toward a Noise Fusion Convolutional Neural Network for Image Denoising, proposes a Noise Fusion Convolutional Neural Network (NFCNN) designed for image denoising. Compared to existing methods, NFCNN demonstrates a notable advantage in preserving texture details. This work has been accepted for publication in Signal, Image and Video Processing, a journal published by Springer.

The offcial full paper is available here.
Apr 30, 2021 We are delighted to share that our paper entitled An Efficient Feature-Preserving Image Denoising Algorithm Based on a Spatial-Fractional Anisotropic Diffusion Equation has been accepted by the East Asian Journal on Applied Mathematics.

Check this page to view the official paper.

selected publications

  1. 2021_nfcnn.png
    NFCNN: Toward a Noise Fusion Convolutional Neural Network for Image Denoising
    Maoyuan Xu and Xiaoping Xie*
    In Signal, Image and Video Processing, 2021
    arXiv
  2. 2020_zoobuilder.png
    ZooBuilder: 2D and 3D Pose Estimation for Quadrupeds Using Synthetic Data
    Abassin Sourou Fangbemi*, Yi Fei Lu, Mao Yuan Xu, Xiao Wu Luo, Alexis Rolland, and Chedy Raissi
    In SCA, 2020
    arXiv Website
  3. 2021_frac_eq_img_denoising.png
    An efficient feature-preserving PDE algorithm for image denoising based on a spatial-fractional anisotropic diffusion equation
    Maoyuan Xu and Xiaoping Xie*
    In East Asian Journal on Applied Mathematics, 2021
    arXiv