A Paper by Doctoral Student Guy Fargion Was Accepted to EUROGRAPHICS 2026

Doctoral student Guy Fargion will be presenting at the acclaimed EUROGRAPHICS international conference on computer graphics, which will be held in May 2026 in Aachen, Germany. Co-written with his supervisor, Prof. Ofir Weber, the paper will also be published in the official journal of the conference, Computer Graphics Forum.

As part of his doctoral research in the computer engineering track, Fargion focuses on digital geometry processing. In the paper, Fargion and Prof. Weber present a new algorithm for solving a fundamental problem in computer graphics: how to flatten 3D surfaces onto the 2D plane – a process known as surface parameterization. "This problem has been extensively studied over the past 30 years, and new algorithms offering increasingly efficient solutions continue to be published annually," says Fargion. "Its popularity stems from a recurring computational need: every time a three-dimensional object is displayed on screen – such as a character in a game, film, or animation – it needs to be colored or textured. To do this, the model must first be 'flattened' into two dimensions, which greatly simplifies the texturing process. The colors and textures are then mapped back onto the surface of the three-dimensional model."

But the transition from three to two dimensions introduces engineering challenges, the most significant of which is geometric distortion. This distortion, Fargion explains, is unavoidable, but it can be minimized. "When such distortions are present, discrepancies arise between the texturing performed in two dimensions and the final result on the three-dimensional model. To minimize these distortions, the most advanced methods, those that yield the highest-quality results, solve complex optimization problems and are therefore quite slow. A popular approach to significantly accelerating this process is to restrict the space of possible solutions, a technique known as the harmonic subspace. While this approach enables fast solution search, it has a key drawback: due to the restriction on the solution space, the result found by the algorithm is often far from optimal and can sometimes be relatively poor."

Fargion and Prof. Weber's paper proposes a solution to this problem. "In our research, we introduced a method for computing a slight modification to the classical harmonic subspace constraints, so that the search remains fast as before, but the solution space now includes options much closer to the optimum that could have been achieved without any constraints. The modification is based on the result obtained by briefly running the slow, high-quality method on a heavily simplified version of the model, which requires minimal computation time."

The results demonstrate that the algorithm proposed by Fargion and Prof. Weber is capable of finding very high-quality solutions – nearly on par with the slowest methods that yield the best results – while maintaining a low overall runtime. This translates to a faster and higher-quality transition from three dimensions to two, with the potential to significantly accelerate graphics rendering in animated films and computer games.