Julia Guerrero Viu defenderá su tesis doctoral titulada “Perceptually Inspired Learning Models for Intuitive Authoring of Material Appearance” el lunes 2 de febrero de 2026 a las 15:00 CET. La defensa se llevará a cabo en la Sala de Conferencias I3A (Planta 1) del edificio I+D+I.
La investigación ha sido dirigida por Belén Masiá Corcoy y Diego Gutiérrez Pérez profesores de nuestro Departamento.
PhD Thesis Defense
Title: Perceptually Inspired Learning Models for Intuitive Authoring of Material Appearance
Author: Julia Guerrero Viu
Supervisors: Belén Masiá Corcoy & Diego Gutiérrez Pérez
Date & Time: Today, February 2, 2026 at 15:00 CET
Place: Sala de Conferencias I3A (Planta 1), edificio I+D+I.
Online link (Google Meet): https://meet.google.com/ghy-eycv-tew
Abstract:
Our visual perception of the world is strongly influenced by how we interpret material appearance. While humans have an innate ability to recognize high-level appearance properties like glossiness or softness, these characteristics emerge from complex interactions between different factors, including not only surface reflectance, but also external conditions such as illumination, geometry, or the subjective prior knowledge of the viewer. In computer graphics, there is still a fundamental gap between computational representations of appearance and these perceptually meaningful, human-friendly properties, limiting our ability to interact effectively with material authoring tools. This thesis explores how human perception can inform the learning of latent representations that are better aligned with how we see and understand material appearance. We demonstrate how these representations can be used to develop more intuitive and controllable authoring tools, such as texture editing, material captioning or fine-grained material selection, ultimately expanding the creative possibilities in visual content creation.
Enlaces:
https://diis.unizar.es/noticias/tesis-julia-guerrero
https://webdiis.unizar.es/~juliagv/
http://webdiis.unizar.es/~bmasia/
http://giga.cps.unizar.es/~diegog/
Publicaciones más importantes:
“The Visual Language of Fabrics” (TOG, SIGGRAPH 2023), https://valentin.deschaintre.fr/text2fabric
