External Research Collaborator @ Meta
Nonstop Consulting, Zurich, Switzerland (Remote)
- Part-time role to continue my work alongside the Monaco rendering team at Meta, working on real-time physically based rendering.
Jorge Condor
PhD Student
I am a PhD Student on Computer Graphics at IDSIA-USI (Lugano, Switzerland) under the supervision of Piotr Didyk. My research focuses on leveraging inverse graphics to improve applications in vision, rendering and fabrication. During my PhD I had the opportunity to intern at Meta Reality Labs working alongside the talented Monaco Rendering Team, working on accelerating physically-based volume rendering for photorealistic avatars, under the supervision of Christophe Hery and Adrián Jarabo. Starting from Summer 2025 I will be joining NVIDIA as a Research Scientist Intern at the Toronto AI Lab, under the supervision of Zan Gojcic. I'm also an avid photographer and I love trekking in my spare time.
- Download my CV here
- Location: Lugano, Switzerland
- Email: jorge.condor@usi.ch
Publications
Puzzle Similarity: A Perceptually-guided No-Reference Metric for Artifact Detection in 3D Scene Reconstructions
ArXiv [Nov'24]
A novel no-reference metric to spatially assess 3D reconstructed view quality using perceptual embeddings, through an efficient patch-wise similarity computation between the training dataset and the evaluated view. Surpasses even direct reference metrics in human assessment correlation tests. We leverage it in automatic, recursive in-painting for artifact restoration.
Perceptually Optimized Super Resolution
ArXiv [Nov'24]
We leverage low-level human vision models to drastically improve the efficiency of learned Super Resolution (SR) models. Our key insight is that compute should be focused on areas that are perceptually significant to the human eye. By parameterizing the upsampling capacity of any off-the-shelf SR model in the frequency domain, we can adaptively upsample content based on the capacity of a human observer to perceive the difference between cheap/simple upsampling (i.e. bicubic interpolation) and expensive learned models, smoothly filling the gap in between through model branching and quantization strategies.
Don't Splat your Gaussians: Volumetric Ray-Traced Primitives for Modeling and Rendering Scattering and Emissive Media
ACM Transactions on Graphics [TOG'24]
Banking on the popularity of rasterized 3D Gaussian Splatting methods, we formalize the ray-tracing of volumes composed of kernel mixture models (Gaussian or otherwise). Our physically-based, path-traced formulation allows us to render and optimize both scattering and emissive volumes, as well as radiance fields, in an extremely efficient and compact manner. We also introduce the Epanechnikov kernel as an efficient alternative for the Gaussian kernel in radiance field rendering, and showcase the advantages of a ray-traced framework, while maintaining real-time performance.
Gloss-aware Color Correction for 3D Printing
Proceedings of SIGGRAPH [SIGG'23]
We measure, model and correct perceived color shifts in 3D printed objects with different levels of gloss. To formulate the correction we conduct novel perceptual experiments and propose an optimization procedure solved using differentiable rendering.
A Learned Radiance-Field Representation for Complex Luminaires
Eurographics Symposium on Rendering [EGSR'22]
An efficient method for rendering complex luminaires using a high quality octree-based representation of the luminaire emission modelled through Neural Radiance Fields, achieving up to x100 speedups over explicit modelling
A Generative Framework for Image-based Editing of Material Appearance using Perceptual Attributes
Computer Graphics Forum, presented at Eurographics 2022 [CGF'21]
An image-based editing method that allows to modify the material appearance of an object by increasing or decreasing high-level perceptual attributes, using a single image as input.
Normal Map Estimator from Single RGB Images of Objects in the Wild
Jornada Jóvenes Investigadores I3A 2021, Poster Session [JJI-I3A'21]
We apply recent advancements in generative neural networks and an untapped dataset to the problem of volume estimation from a single RGB input and train a model that performs the task efficiently, ready to deploy in other vision and computer graphics pipelines
Experience
Professional Experience
Research Scientist Intern
Meta Reality Labs, Zurich, Switzerland
- Worked with the Monaco rendering team on real-time solutions for Physically-based rendering of humans involving hardware-accelerated analytic ray tracing, Gaussian splatting and differentiable rendering, under the supervision of Christophe Hery and Adrián Jarabo
Research Intern
Graphics and Imaging Lab, Zaragoza, Spain
- Funded by a competitive scholarship granted by the I3A (Instituto de Investigación en Ingeniería de Aragón) rewarding excellent academic records
- Developed a volume estimation method from single RGB images based on generative AI
- Colaborated on a framework for image-based editing of material appearance using perceptual attributes
- Extended Neural Radiance Field methods to accelerate traditional rendering pipelines in the context of rendering of complex luminaires
Private Tutor
Zaragoza, Spain (own account)
- Mathematics, Physics and Chemistry tutor for baccalaureate (university entry exams' preparation) students
Teaching & Supervision
Teaching Assistant, Digital Image Processing and Graphics
USI Lugano, Switzerland
Development of assignments and exams, grading, onsite teaching of exercise lectures, student tutoring on BSc and MSc courses on digital image and video processing and computer graphics
Thesis and Research Supervisor
USI Lugano, Switzerland
Research Interns
USI Lugano, Switzerland
- Xiana Carrera Alonso [July 2023 - Sept 2023]
- Arnaud Fauconnet [July 2023 - Sept 2023]
Master Thesis
USI Lugano, Switzerland
- Michele Chersich, Masters in Computational Science [Feb 2023 - September 2024]
- Nicolai Hermann, Masters in AI [Feb 2024 - September 2024]
- Kacper Kramarz-Fernandez, Masters in High Performance Computing (EU4HPC Programme) [November 2024 - Ongoing]
Reinforcement Learning for Importance Sampling of Neural Radiance Fields
Perceptually-Driven Neural Inpainting for Seamless 3D Reconstructions
Efficient Physically-Based Rendering of Kernel-based Volumes
Electronics Maturity Projects for Highschool Students
Liceo Cantonale 2, Lugano, Switzerland
Development of an IoT device based on LoRA and ESP32 for plant health monitoring and control with highschool students as part of their Maturity Projects (Lavoro di Maturità or LaM).
Education
PhD in Computer Science
IDSIA-USI Lugano, Switzerland
PhD Candidate on Computer Graphics under the supervision of Prof. Piotr Didyk. My research focuses on leveraging inverse graphics to improve applications in vision, graphics and fabrication.
Master in Graphics, Robotics and Computer Vision
Universidad de Zaragoza, Spain
Highly competitive Master with a strong research focus. Took courses on Deep Learning, Computer Graphics, VR, Computer Vision, SLAM and Robotics. Obtained Honors in Modelling and Simulation of Appearance (Computer Graphics course), where I developed a path tracer based on Nori (educational Mitsuba) and implemented many features including a full volumetric path tracer for both homogeneous and heterogeneous media and fur rendering. Got Honorable Mention (second prize) in the Rendering Contest judged by Marcos Fajardo, Matt Chiang, and Wojciech Jarosz (check my submission here )
University was Ranked #1 in Spain in Computer Graphics, Robotics and Computer Vision in 2020-2022(edurank)
BSc, MSc Courses on Electrical Engineering (Erasmus)
Aalto University, Helsinki, Finland
Took Master-level courses in the fields of AI, electronics design and robotics, working in a highly cooperative and diverse environment
Bachelor in Electronics and Control Engineering
Universidad de Zaragoza, Spain
Top 7% of my promotion. Special interest in digital and analog electronics, robotics and machine learning. Class delegate for several years. Obtained Honors in Digital Electronics, Thermodynamics, Chemistry and Fundamentals of Electronics
