Synthetic Data Generation for Computer Vision
At Indra, I led a 3D simulation and rendering project that generated realistic synthetic training data for computer-vision detection models, as part of autonomous systems for air-traffic-control operations.
The idea is to render large, automatically-labelled image sets directly from 3D models: orbiting a virtual camera around each object while varying pose, lighting, and physics (including soft-body deformation), so a detector can be trained without any hand-annotated real imagery. The open-source Blender tool I built captures this pipeline, producing multi-view renders and segmentation-ready data from arbitrary 3D assets.
The hard part is the synthetic-to-real gap: a detector trained only on clean renders usually loses accuracy on real footage, because synthetic textures, lighting, sensor noise, and backgrounds never quite match the real world. Closing that gap is most of the work. By pushing domain randomization, deliberately varying materials, lighting, camera parameters, and physics across the dataset, the model is forced to rely on features that actually transfer to reality instead of overfitting to the look of the simulator, so it keeps its accuracy when it finally sees real images.
I am a PhD student in Computer Vision at the Institute of Robotics and Industrial Informatics (IRI, CSIC–UPC) in Barcelona, advised by Antonio Agudo. My work sits at the intersection of computer vision and computer graphics, where I study how to turn ordinary images and video into faithful, editable 3D representations of the real world.
Broadly, I am interested in 3D scene understanding: reconstructing scenes, recognizing and segmenting the objects within them, and decomposing them into clean, manipulable parts that can be edited, relit, or reused. I work with mesh and splatting representations, multi-view segmentation, and generative models for 3D content, always with an eye toward connecting what a camera sees to the underlying structure of a scene.
Before starting my PhD I worked as an AI engineer and as a research assistant, on projects spanning 3D simulation, synthetic data generation, and interactive graphics. I also teach as a teaching assistant at UPF, for the undergraduate Computer Graphics and Electronic Games courses, the latter where students learn to code their own videogame.
Outside research, I am a videogame and football lover and an amateur weightlifter, and a fan of self-hosting, Linux, and open-source software. I enjoy building systems as much as studying them, and I am always happy to discuss research, 3D vision and graphics, or potential collaborations.