Matheus Medeiros de Souza

I am a final-year Ph.D. student at KAUST Computational Imaging Group, working with Prof. Wolfgang Heidrich.

My research is centered on the innovative fusion of deep reconstruction algorithms and end-to-end optics design. This involves developing deep learning models and formulating optimization strategies that incorporate optical elements into the computational loop. Focusing on the following topics:

  • Minimalistic camera design.
  • Deep learning for optical design.
  • Computational cameras.
  • Optics-aware computational photography.

    • Email  /  Google Scholar  /  LinkedIn

    		profile photo
    Education

    Image

    Ph.D. in Computer Science

    King Abdullah University of Science and Technology (KAUST), Saudi Arabia

    2022 - Present
    Image

    M.Sc. in Computer Science

    King Abdullah University of Science and Technology (KAUST), Saudi Arabia

    2020 - 2022
    Image

    B.Sc. in Electrical Engineering

    Universidade Federal do Rio Grande do Norte (UFRN), Brazil

    2015 - 2019

    Graduated with Highest Honors • Best Student Award, Class of 2019
    Work Experience

    Image

    Research Scientist Intern in Computational Imaging

    Meta Reality Labs - Surreal Team

    Mentored by Michael Goesele

    June 2025 - November 2025
    Publications
    Image
    Image
    		Low Latency Gaze Tracking via Latent Optical Sensing
    Yidan Zheng*, Matheus Souza*, Kaizhang Kang, Qiang Fu, Hadi Amata, Wolfgang Heidrich
    *equal contribution

    SIGGRAPH Asia 2026

  • Real-time gaze tracking with passive optical encoding via a microlens array and co-designed binary mask.
  • A 4×4 phototransistor array with a lightweight network achieves 3.4 ms end-to-end sensing-to-inference latency.

    		•
    Image
    Image
    		Latent Space Imaging
    Matheus Souza, Yidan Zheng, Kaizhang Kang, Yogeshwar Nath Mishra, Qiang Fu, Wolfgang Heidrich

    CVPR 2025, Code

  • New paradigm for very low bandwidth image capture based on the latent space of generative models.
  • The demonstration of a range of downstream applications on this latent space with real hardware experimentation.

    		•
    Image
    Image
    		Limitations of Data-Driven Spectral Reconstruction - An Optics-Aware Analysis
    Qiang Fu*, Matheus Souza*, Suhyun Shin, Eunsue Choi, Seung-Hwan Baek, Wolfgang Heidrich
    *equal contribution

    Computational Optical Sensing and Imaging, 2024. Oral Presentation
    IEEE TCI

  • Comprehensive analysis of atypical overfitting in state-of-the-art data-driven hyperspectral imaging.
  • Optical aberrations can provide encoding power if modeled correctly.

    		•
    Image
    Image
    		End-to-End Hybrid Refractive-Diffractive Lens Design with Differentiable Ray-Wave Model
    Xinge Yang, Matheus Souza, Kunyi Wang, Praneeth Chakravarthula, Qiang Fu, Wolfgang Heidrich

    SIGGRAPH Asia 2024. Paper (Arxiv) / Paper (PDF) / Supp (PDF)

  • Differentiable ray-tracing and wave-propagation model.
  • End-to-End hybrid refractive-diffractive lenses design with prototypes.

    		•
    Image
    Image
    		MetaISP - Exploiting Global Scene Structure for Accurate Multi-Device Color Rendition
    Matheus Souza, Wolfgang Heidrich

    MetaISP VMV 2023 / Code
    CRISPnet: Color rendition ISP net. Paper (Arxiv)

  • We developed a model for learning multiple commercial ISPs.
  • Integrating global scene semantics, metadata information, and advanced deep learning techniques.
  • Collected synthetic and real-world datasets, consisting of RAW-RGB pairs from various devices.

    		•

    The website template is from Dr. John Barron.