Lumirithmic Ltd

Camera intelligence, grounded in the physics of light.

Lumirithmic is a UK company spun out of Imperial College London, working at the intersection of computer vision, computer graphics, computational imaging, and AI. We turn camera observations into useful representations of shape, appearance, materials, and human faces, and develop imaging systems, custom models, and production software.

Our research connects practical capture hardware with physically based modelling and learned representations. It spans high-quality facial capture using commodity displays and cameras, neural inverse rendering and relighting, spectral skin reflectance, biophysical appearance and ageing, polarization imaging, and material acquisition.

Website · About us · Work with us · LinkedIn

Research areas

• Digital humans: facial geometry, skin appearance, photometric normals, and realistic avatars.
• Inverse rendering and relighting: separating illumination from reflectance and recovering reusable appearance representations.
• Spectral and biophysical imaging: skin chromophores, fluorescence, blood distributions, and wavelength-dependent appearance.
• Computational capture: controlled illumination, polarization, desktop displays, and smartphone imaging.
• Materials and generative graphics: reflectance acquisition, spectral rendering, and physically grounded AI.

Our research & publications

Lumirithmic develops and supports research that advances how we capture, understand, and reproduce the visual world. This showcase brings together publications from our researchers and the work we have built with our academic partners — from practical facial capture and skin imaging to neural rendering, materials, and spectral graphics.

Our university collaborations include Imperial College London, the University of York, and NTNU, alongside other institutions credited in the individual papers. Explore the publications below for the science behind our technology.

2026

• Spectral Rendering Without a Spectral Renderer: Learned Spectral Codes for RGB Pipelines
  Jiaqi Yu, Dar’ya Guarnera, Giuseppe Claudio Guarnera. Accepted at SIGGRAPH Asia 2026; to appear in ACM Transactions on Graphics 45(6). Learned spectral representations for use with standard RGB renderers. Earlier preprint title: Compact Hadamard Latent Codes for Efficient Spectral Rendering.

2025

• Single-Shot Facial Capture using Polarized RGB Sinusoidal Illumination
  Arvin Lin, Abhijeet Ghosh. SIGGRAPH Asia 2025 Conference Papers. Facial reflectance and photometric normals from a single exposure per viewpoint. DOI

• Lighting Beneath the Surface: Spectral and Polarised Cues for Next-Generation Facial Imaging
  Giuseppe Claudio Guarnera, Dar’ya Guarnera, Xiaohui Li. London Imaging Meeting 2025 / Journal of Physics: Conference Series 3128, 012013. Spectral and polarization cues for facial appearance, re-ageing, and geometry. DOI

• Image Adaptation for Colour Vision Deficient Viewers using Vision Transformers
  Tom Gillooly, Jean-Baptiste Thomas, Jon Y. Hardeberg, Giuseppe Claudio Guarnera. WACV 2025. Attention-guided image recolouring for colour vision deficiency.

2024

• High-Quality Facial Geometry from Sparse Heterogeneous Cameras under Active Illumination
  Lewis Bridgeman, Gilles Rainer, Abhijeet Ghosh. CVMP 2024. Neural surface reconstruction for sparse facial capture systems.

• Practical RGB Measurement of Fluorescence and Blood Distributions in Skin
  Emilie Nogué, Arvin Lin, Xiaohui Li, Giuseppe Claudio Guarnera, Abhijeet Ghosh. Color and Imaging Conference 2024. Estimating skin fluorescence and blood distributions using RGB imaging and display illumination. DOI

• Realistic Facial Age Transformation with 3D Uplifting
  Xiaohui Li, Giuseppe Claudio Guarnera, Arvin Lin, Abhijeet Ghosh. Computer Graphics Forum / EGSR 2024. Joint transformation of facial appearance and 3D shape with biophysical skin edits.

• High Quality Neural Relighting using Practical Zonal Illumination
  Arvin Lin, Yiming Lin, Xiaohui Li, Abhijeet Ghosh. EGSR 2024. Neural relighting from practical monitor-based capture. DOI

• Practical Measurement and Neural Encoding of Hyperspectral Skin Reflectance
  Xiaohui Li, Giuseppe Claudio Guarnera, Arvin Lin, Abhijeet Ghosh. 3DV 2024. Practical skin measurement and neural encoding of hyperspectral appearance.

2023

• Neural Shading Fields for Efficient Facial Inverse Rendering
  Gilles Rainer, Lewis Bridgeman, Abhijeet Ghosh. Computer Graphics Forum / Pacific Graphics 2023. Recovering facial reflectance from photographs under unknown illumination. DOI

• Deep Shape and SVBRDF Estimation using Smartphone Multi-lens Imaging
  Chongrui Fan, Yiming Lin, Abhijeet Ghosh. Computer Graphics Forum / Pacific Graphics 2023. Shape and spatially varying reflectance estimation from simultaneous smartphone images.

• Practical Acquisition of Shape and Plausible Appearance of Reflective and Translucent Objects
  Arvin Lin, Yiming Lin, Abhijeet Ghosh. Computer Graphics Forum / EGSR 2023. Display-based capture of challenging reflective and translucent objects. DOI

2022

• Practical and Scalable Desktop-based High-Quality Facial Capture
  Alexandros Lattas, Yiming Lin, Jayanth Kannan, Ekin Ozturk, Luca Filipi, Giuseppe Claudio Guarnera, Gaurav Chawla, Abhijeet Ghosh. ECCV 2022 — oral presentation. High-quality geometry and appearance capture with commodity displays and cameras.

• Spectral Upsampling Approaches for RGB Illumination
  Giuseppe Claudio Guarnera, Yuliya Gitlina, Valentin Deschaintre, Abhijeet Ghosh. EGSR 2022. Recovering spectral illumination representations from RGB inputs.

• Polarization-imaging Surface Reflectometry using Near-field Display
  Emilie Nogué, Yiming Lin, Abhijeet Ghosh. EGSR 2022. Practical acquisition of spatially varying material appearance with polarization and display illumination.

Talks, demos, and workshop abstracts

• Desktop-based High-Quality Facial Capture for Everyone — SIGGRAPH 2022 Talks. Alexandros Lattas and colleagues. DOI
• Lumirithmic Demo: Desktop-based High-Quality Facial Capture — CVMP 2022 demo, included in the conference programme.
• A neural approach for spectral reconstruction of facial RGB images — APPAMAT/IS&T International Workshop on Material Appearance 2023, abstract. Fereshteh Mirjalili, Jon Yngve Hardeberg, Giuseppe Claudio Guarnera.

More from our scientific team

Our scientific team also advances the wider field through academic collaborations. The following publications are co-authored by our Chief Science Officer, Giuseppe Claudio Guarnera, and complement our work in material appearance, controllable generation, and spectral imaging.

• PBR-Inspired Controllable Diffusion for Image Generation — Bowen Xue, Giuseppe Claudio Guarnera, Shuang Zhao, Zahra Montazeri. Computer Graphics Forum, 2026. Preprint
• Physics-Guided Motion Loss for Video Generation Model — Bowen Xue, Giuseppe Claudio Guarnera, Shuang Zhao, Zahra Montazeri. arXiv preprint, 2025.
• ReflectanceFusion: Diffusion-based text to SVBRDF Generation — Bowen Xue, Giuseppe Claudio Guarnera, Shuang Zhao, Zahra Montazeri. EGSR 2024. Preprint
• A Neural Approach for Skin Spectral Reconstruction — Fereshteh Mirjalili, Giuseppe Claudio Guarnera. London Imaging Meeting 2024. Paper

Collaborate with Lumirithmic

We work with research and industry partners on imaging systems, facial and material appearance, custom computer vision models, and production integration. Get in touch to discuss a research collaboration or a camera AI project.

---

Publication list updated 14 October 2026.