Why does It Look Different? - Material Appearance International Doctoral Network
Human evolution has endowed us with refined visual and cognitive skills, enabling material appearance assessment across diverse conditions and contexts. While the physics of light transport within materials is well understood, the visual mechanisms underlying perceptual equivalence among materials remain largely unexplored, and no reliable metric currently quantifies perceptual differences between them. This gap is a critical barrier to advancing aesthetic product design and manufacturing, particularly with the advent of complex designs featuring visual (color, gloss, translucency, texture) and haptic (softness, roughness) attributes made possible by new manufacturing technologies like multimaterial 3D printing. These innovations facilitate decentralized production, reducing supply chain dependence and energy consumption while enabling high-quality, customized parts, such as aesthetic dental restorations. However, current quality control methods rely on subjective visual inspection, creating variability and limiting reproducibility, especially for products requiring high appearance accuracy. To overcome this, appearance difference metrics are essential for quality control and design adjustments in line with manufacturing constraints (Appearance Gamut Mapping). Furthermore, generative AI (genAI) models in automatic product design require differentiable loss functions that capture human perception of material appearance. The WILD MAIDN project aims to establish foundational metrics for appearance difference and equivalence, advancing visual perception science and providing tools for automated, perceptually-driven design and production that can ultimately support future genAI applications.
4/6 complete
Imported from OpenAIRE Germany | Grant 101312447.
Open sourceOverview
Grant agreement 101312447 | WILD-MAIDN
Open sourceHuman evolution has endowed us with refined visual and cognitive skills, enabling material appearance assessment across diverse conditions and contexts. While the physics of light transport within materials is well understood, the visual mechanisms underlying perceptual equivalence among materials remain largely unexplored, and no reliable metric currently quantifies perceptual differences between them. This gap is a critical barrier to advancing aesthetic product design and manufacturing, particularly with the advent of complex designs featuring visual (color, gloss, translucency, texture) and haptic (softness, roughness) attributes made possible by new manufacturing technologies like multimaterial 3D printing. These innovations facilitate decentralized production, reducing supply chain dependence and energy consumption while enabling high-quality, customized parts, such as aesthetic dental restorations. However, current quality control methods rely on subjective visual inspection, creating variability and limiting reproducibility, especially for products requiring high appearance accuracy. To overcome this, appearance difference metrics are essential for quality control and design adjustments in line with manufacturing constraints (Appearance Gamut Mapping). Furthermore, generative AI (genAI) models in automatic product design require differentiable loss functions that capture human perception of material appearance. The WILD MAIDN project aims to establish foundational metrics for appearance difference and equivalence, advancing visual perception science and providing tools for automated, perceptually-driven design and production that can ultimately support future genAI applications.