EG 2026 - Short Papers

Permanent URI for this collection

EG 2026 Short Papers
Faces, Characters & Human Modeling
ARTIST: Adaptive Humanoid Rigging by Transferring Individual Style with Optimal Transport
Lefèvre Jeanne-Emma, Cheynel Théo, El Khalifi Omar, Daniel Thomas, Bellot-Gurlet Baptiste
Skeleton Subspace Skin Penetration Removal
Mukai Tomohiko, Taketomi Takafumi
Vid2Haircut: 3D Strand-Based Hairstyle Reconstruction from Video
Ben Ayed Fatma, Becherini Giorgio, Thies Justus, Sklyarova Vanessa
StyleYourSmile: Diffusion-Driven One Shot Cross-Domain Retargeting for Portraits
Dey Avirup, Namboodiri Vinay
Token-Based Dual-Codebook Learning for Robust 3D Pose Lifting
Jeon Minsu, Kim Janghyun, Park Jinsun
RAW: Robust AvatarWatermarking - Benchmarking and Baseline
Parry Jack, Saunders Jack, Namboodiri Vinay
Beyond FID: Human Perceptual Judgments Reveal Systematic Blind Spots in GAN Face Evaluation
Nierula Birgit, Melnik Anna, Barthel Florian, Brama Aileen, Hilsmann Anna, Eisert Peter, Nikulin Vadim V., Gaebler Michael, Klotzsche Felix, Chen Yonghao, Stephani Tilman, Bosse Sebastian
Appearance, Imaging & Tools
RetiDiff: Stable Underwater Image Color Reconstruction Based on Retinex and Diffusion Distillation
Qiu Wenyao, Zhou Zhuang, Zhang Xin, Chen Jiayi, Zhou Shiping, Tao Ran
A Delaunay Keyer for Colorspace-Local Matte Extraction and Spill Suppression
Criddle Isaac, Holladay Seth, Egbert Parris
Controllable Cinemagraph Generation from A Still Image
Van Thanh Le, Ito Daichi, Mahapatra Aniruddha, Mai Long, Singh Krishna Kumar, Kulkarni Kuldeep, Liu Feng, Fu Yun, Yoon Jae Shin
On the Accuracy of Surface Scattering Theories
Avolio Matthew, D’Eon Eugene, Steinberg Shlomi
On Cosine Prior Distributions for Neural Path Guiding
Gutsch Jan-Luca, Dereviannykh Mikhail, Hanika Johannes
Implementation is Illustration: Zero-Overhead On-Demand Visualization for High-Performance Linear Algebra
Rautek Peter
Capsule: Efficient Player Isolation for Datacenters
Du Zhouheng, Davari Nima, Li Li, Loi Wei Sen, Kodirov Nodir
Simulation, Geometry & Computational Design
VisACD: Visibility-Based GPU-Accelerated Approximate Convex Decomposition
Fokin Egor, Savva Manolis
Tetrahedron-Tetrahedron Intersection and Volume Computation Using Neural Networks
Erendiro Pedro, Meißenhelter Hermann, Zachmann Gabriel
ConJEB: A Large Elastic Contact Jet Engine Bracket Quadratic Program Dataset
Ferreira Stephanie, Giebel Andreas, Mueller-Roemer Johannes Sebastian
Computational Design of Forced-Perspective Structures
Watanabe Sarika, Fukusato Tsukasa
What a ComfortableWorld: Ergonomic Principles Guided Apartment Layout Generation
Nieciecki Piotr, Plocharski Aleksander, Musialski Przemyslaw
Beyond Segmentation: Structurally Informed Facade Parsing from Imperfect Images
Janicki Maciej, Plocharski Aleksander, Musialski Przemyslaw
Differentiable Objectives for 3D Scene Relighting via Gradient Descent on OLAT Basis Coefficients
Savage Anson, Egbert Parris, Holladay Seth
Rendering Representations & GPU Pipelines
2D-SuGaR: Surface-Aware Gaussian Splatting for Geometrically Accurate Mesh Reconstruction
Gupta C. R. Prajwal, Sheth Divyam, Ha Jinjoo, Ostrek Mirela, Thies Justus
OT-UVGS: Revisiting UV Mapping for Gaussian Splatting as a Capacity Allocation Problem
Kim Byunghyun
VkSplat: High-Performance 3DGS Training in Vulkan Compute
Chen Jingxiang, Ibrahim Mohamed, Liu Yang
RETA3D: Real-Time Animatable 3D Gaussian Head Generation
Chen Shu-Yu, Qiu Chunshuo, Liu Feng-Lin, Cao Yanpei, Fu Hongbo, Gao Lin
Voxel Deformation-Aware Neural Intersection Function
Kao Chih-Chen, Makowski Grzegorz, Fujieda Shin, Harada Takahiro
Robust Ray–Surface Intersections for Algebraic Surfaces
Szente Péter, Karikó Csongor Csanád, Valasek Gábor
Helper-Lane Optimized Triangulation of Polygons
Bene Róbert, Valasek Gábor

BibTeX (EG 2026 - Short Papers)
@inproceedings{
10.2312:egs.20261001,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
ARTIST: Adaptive Humanoid Rigging by Transferring Individual Style with Optimal Transport}},
author = {
Lefèvre, Jeanne-Emma
and
Cheynel, Théo
and
El Khalifi, Omar
and
Daniel, Thomas
and
Bellot-Gurlet, Baptiste
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261001}
}
@inproceedings{
10.2312:egs.20261002,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
Skeleton Subspace Skin Penetration Removal}},
author = {
Mukai, Tomohiko
and
Taketomi, Takafumi
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261002}
}
@inproceedings{
10.2312:egs.20261003,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
Vid2Haircut: 3D Strand-Based Hairstyle Reconstruction from Video}},
author = {
Ben Ayed, Fatma
and
Becherini, Giorgio
and
Thies, Justus
and
Sklyarova, Vanessa
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261003}
}
@inproceedings{
10.2312:egs.20261004,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
StyleYourSmile: Diffusion-Driven One Shot Cross-Domain Retargeting for Portraits}},
author = {
Dey, Avirup
and
Namboodiri, Vinay
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261004}
}
@inproceedings{
10.2312:egs.20261005,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
Token-Based Dual-Codebook Learning for Robust 3D Pose Lifting}},
author = {
Jeon, Minsu
and
Kim, Janghyun
and
Park, Jinsun
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261005}
}
@inproceedings{
10.2312:egs.20261006,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
RAW: Robust AvatarWatermarking - Benchmarking and Baseline}},
author = {
Parry, Jack
and
Saunders, Jack
and
Namboodiri, Vinay
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261006}
}
@inproceedings{
10.2312:egs.20261007,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
Beyond FID: Human Perceptual Judgments Reveal Systematic Blind Spots in GAN Face Evaluation}},
author = {
Nierula, Birgit
and
Melnik, Anna
and
Stephani, Tilman
and
Bosse, Sebastian
and
Barthel, Florian
and
Brama, Aileen
and
Hilsmann, Anna
and
Eisert, Peter
and
Nikulin, Vadim V.
and
Gaebler, Michael
and
Klotzsche, Felix
and
Chen, Yonghao
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261007}
}
@inproceedings{
10.2312:egs.20261008,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
RetiDiff: Stable Underwater Image Color Reconstruction Based on Retinex and Diffusion Distillation}},
author = {
Qiu, Wenyao
and
Zhou, Zhuang
and
Zhang, Xin
and
Chen, Jiayi
and
Zhou, Shiping
and
Tao, Ran
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261008}
}
@inproceedings{
10.2312:egs.20261010,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
Controllable Cinemagraph Generation from A Still Image}},
author = {
Le, Van Thanh
and
Ito, Daichi
and
Mahapatra, Aniruddha
and
Mai, Long
and
Singh, Krishna Kumar
and
Kulkarni, Kuldeep
and
Liu, Feng
and
Fu, Yun
and
Yoon, Jae Shin
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261010}
}
@inproceedings{
10.2312:egs.20261011,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
On the Accuracy of Surface Scattering Theories}},
author = {
Avolio, Matthew
and
D'Eon, Eugene
and
Steinberg, Shlomi
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261011}
}
@inproceedings{
10.2312:egs.20261012,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
On Cosine Prior Distributions for Neural Path Guiding}},
author = {
Gutsch, Jan-Luca
and
Dereviannykh, Mikhail
and
Hanika, Johannes
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261012}
}
@inproceedings{
10.2312:egs.20261009,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
A Delaunay Keyer for Colorspace-Local Matte Extraction and Spill Suppression}},
author = {
Criddle, Isaac
and
Holladay, Seth
and
Egbert, Parris
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261009}
}
@inproceedings{
10.2312:egs.20261013,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
Implementation is Illustration: Zero-Overhead On-Demand Visualization for High-Performance Linear Algebra}},
author = {
Rautek, Peter
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261013}
}
@inproceedings{
10.2312:egs.20261014,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
Capsule: Efficient Player Isolation for Datacenters}},
author = {
Du, Zhouheng
and
Davari, Nima
and
Li, Li
and
Loi, Wei Sen
and
Kodirov, Nodir
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261014}
}
@inproceedings{
10.2312:egs.20261015,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
VisACD: Visibility-Based GPU-Accelerated Approximate Convex Decomposition}},
author = {
Fokin, Egor
and
Savva, Manolis
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261015}
}
@inproceedings{
10.2312:egs.20261016,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
Tetrahedron-Tetrahedron Intersection and Volume Computation Using Neural Networks}},
author = {
Pedro, Erendiro
and
Meißenhelter, Hermann
and
Zachmann, Gabriel
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261016}
}
@inproceedings{
10.2312:egs.20261017,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {
Musialski, Przemyslaw
and
Lim, Isaak
}, title = {{
ConJEB: A Large Elastic Contact Jet Engine Bracket Quadratic Program Dataset}},
author = {
Ferreira, Stephanie
and
Giebel, Andreas
and
Mueller-Roemer, Johannes Sebastian
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261017}
}
@inproceedings{
10.2312:egs.20261018,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
Computational Design of Forced-Perspective Structures}},
author = {
Watanabe, Sarika
and
Fukusato, Tsukasa
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261018}
}
@inproceedings{
10.2312:egs.20261019,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
What a ComfortableWorld: Ergonomic Principles Guided Apartment Layout Generation}},
author = {
Nieciecki, Piotr
and
Plocharski, Aleksander
and
Musialski, Przemyslaw
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261019}
}
@inproceedings{
10.2312:egs.20261020,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
Beyond Segmentation: Structurally Informed Facade Parsing from Imperfect Images}},
author = {
Janicki, Maciej
and
Plocharski, Aleksander
and
Musialski, Przemyslaw
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261020}
}
@inproceedings{
10.2312:egs.20261021,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
Differentiable Objectives for 3D Scene Relighting via Gradient Descent on OLAT Basis Coefficients}},
author = {
Savage, Anson
and
Egbert, Parris
and
Holladay, Seth
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261021}
}
@inproceedings{
10.2312:egs.20261022,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
2D-SuGaR: Surface-Aware Gaussian Splatting for Geometrically Accurate Mesh Reconstruction}},
author = {
Gupta, C. R. Prajwal
and
Sheth, Divyam
and
Ha, Jinjoo
and
Ostrek, Mirela
and
Thies, Justus
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261022}
}
@inproceedings{
10.2312:egs.20261023,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
OT-UVGS: Revisiting UV Mapping for Gaussian Splatting as a Capacity Allocation Problem}},
author = {
Kim, Byunghyun
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261023}
}
@inproceedings{
10.2312:egs.20261024,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
VkSplat: High-Performance 3DGS Training in Vulkan Compute}},
author = {
Chen, Jingxiang
and
Ibrahim, Mohamed
and
Liu, Yang
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261024}
}
@inproceedings{
10.2312:egs.20261025,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
RETA3D: Real-Time Animatable 3D Gaussian Head Generation}},
author = {
Chen, Shu-Yu
and
Qiu, Chunshuo
and
Liu, Feng-Lin
and
Cao, Yanpei
and
Fu, Hongbo
and
Gao, Lin
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261025}
}
@inproceedings{
10.2312:egs.20261026,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
Voxel Deformation-Aware Neural Intersection Function}},
author = {
Kao, Chih-Chen
and
Makowski, Grzegorz
and
Fujieda, Shin
and
Harada, Takahiro
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261026}
}
@inproceedings{
10.2312:egs.20261027,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
Robust Ray–Surface Intersections for Algebraic Surfaces}},
author = {
Szente, Péter
and
Karikó, Csongor Csanád
and
Valasek, Gábor
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261027}
}
@inproceedings{
10.2312:egs.20261028,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
Helper-Lane Optimized Triangulation of Polygons}},
author = {
Bene, Róbert
and
Valasek, Gábor
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20261028}
}
@inproceedings{
10.2312:egs.20262000,
booktitle = {
Eurographics 2026 - Short Papers},
editor = {}, title = {{
EUROGRAPHICS 2026: Short Papers Frontmatter}},
author = {
Lim, Isaak
and
Musialski, Przemyslaw
}, year = {
2026},
publisher = {
The Eurographics Association
},
ISSN = {2309-5059},
ISBN = {978-3-03868-299-8},
DOI = {
10.2312/egs.20262000}
}

Browse

Recent Submissions

Now showing 1 - 29 of 29
  • Item
    ARTIST: Adaptive Humanoid Rigging by Transferring Individual Style with Optimal Transport
    (The Eurographics Association, 2026) Lefèvre, Jeanne-Emma; Cheynel, Théo; El Khalifi, Omar; Daniel, Thomas; Bellot-Gurlet, Baptiste; Musialski, Przemyslaw; Lim, Isaak
    Automatic rigging transforms static meshes into articulated characters by predicting skeletal structure. However, rigging is inherently subjective: artists develop personal preferences for joint placement. Current approaches omit this aspect, learning only the average “style” of their training data. We quantify inter-artist variance through a user study and dataset analysis, demonstrating this notion of “rigging style”. We propose a voxel-based model leveraging pretrained 3D backbones that outperforms state-of-the-art methods. We also introduce a one-shot style adaptation method based on volumetric optimal transport: given a single artist-rigged example, we transfer its stylistic joint placements to any new character. This improves any rigging model and supports different bone counts or hierarchies, reconciling automatic rigging with artistic variability.
  • Item
    Skeleton Subspace Skin Penetration Removal
    (The Eurographics Association, 2026) Mukai, Tomohiko; Taketomi, Takafumi; Musialski, Przemyslaw; Lim, Isaak
    This paper presents a kinematics-based method for removing penetration in skinned models based on skeleton subspace deformation. The penetration removal process is formulated as a multi-target inverse kinematics problem, in which penetrated vertices are offset by adjusting the skeletal pose. To ensure numerical stability, penetrated vertices are dynamically clustered based on deformation gradient similarity, reducing the dimensionality of the IK optimization. Experimental results demonstrate that the proposed method robustly eliminates complex penetration in articulated models while achieving interactive performance.
  • Item
    Vid2Haircut: 3D Strand-Based Hairstyle Reconstruction from Video
    (The Eurographics Association, 2026) Ben Ayed, Fatma; Becherini, Giorgio; Thies, Justus; Sklyarova, Vanessa; Musialski, Przemyslaw; Lim, Isaak
    We present Vid2Haircut, a novel approach for strand-based 3D hair reconstruction from monocular head-motion videos. While existing multi-view methods achieve high-fidelity results, they require controlled capture setups. In contrast, single-image approaches suffer from occlusion-driven ambiguities, particularly in unseen regions such as the back of the head. Recent monocular video methods improve reconstruction by leveraging learned priors, but may struggle under natural head motion. To address this, our approach reconstructs accurate geometry from a short monocular video by leveraging viewpoint variations induced by natural head motion to resolve occlusions. Specifically, we extend a learned hair prior [SZP∗25] by jointly optimizing a shared, scalp-aligned hair map in a canonical space across multiple key-frames. To accommodate hair motion during capture, we incorporate a deformation MLP that predicts residual strand offsets, preventing frame-specific deformations from corrupting the canonical hairstyle. Additionally, we stabilize the reconstruction of poorly observed regions using visibility-aware updates and neighboring-strand smoothness constraints. Experiments on synthetic and real data demonstrate improved back-view consistency, scalp attachment, and overall reconstruction quality compared to state-of-the-art baselines, while requiring only casual monocular video as input.
  • Item
    StyleYourSmile: Diffusion-Driven One Shot Cross-Domain Retargeting for Portraits
    (The Eurographics Association, 2026) Dey, Avirup; Namboodiri, Vinay; Musialski, Przemyslaw; Lim, Isaak
    Cross-domain portrait retargeting requires disentangled control over identity, expressions, and domain-specific stylistic attributes. Existing methods, typically trained on subjects in a single domain, either fail to generalize across image styles, need test-time optimizations, or require fine-tuning with curated multi-style data to achieve domain-invariant identity representations. In this work, we introduce StyleYourSmile, a novel one-shot cross-domain face retargeting method that eliminates these bottlenecks. We propose a dual-encoder architecture alongside an efficient data augmentation strategy for representing domain-invariant identity cues and capturing domain-specific stylistic variations. Leveraging these disentangled control signals, we condition a diffusion model to retarget facial expressions across domains. Extensive experiments demonstrate that StyleYourSmile achieves superior identity preservation and retargeting fidelity across a wide range of visual styles.
  • Item
    Token-Based Dual-Codebook Learning for Robust 3D Pose Lifting
    (The Eurographics Association, 2026) Jeon, Minsu; Kim, Janghyun; Park, Jinsun; Musialski, Przemyslaw; Lim, Isaak
    3D human pose estimation from monocular images is inherently challenging due to frequent occlusions, which introduce significant ambiguity in joint visibility. For instance, regression-based methods are highly sensitive to these ambiguities, often leading to unstable and jittery pose estimates. To overcome these limitations, recent token-based methods discretize poses into structured representations and better capture joint dependencies. However, most existing approaches operate in a frame-wise manner, neglecting temporal continuity and consequently suffering from time-inconsistent predictions. Therefore, we propose a spatio-temporal token-based framework for 3D human pose estimation that explicitly models both spatial and temporal dependencies. In specific, a spatial and temporal tokenizer decomposes 3D pose sequences into discrete spatial and temporal tokens via a dual-codebook design. To predict these tokens from 2D pose sequences, we further develop spatial and temporal token classifiers based on a SemGCN–GraphGRU architecture, enabling effective temporal reasoning while preserving skeletal structure. Extensive experiments on the Human3.6M dataset demonstrate that our method achieves state-of-the-art performance among short-sequence methods, while significantly reducing high-frequency jitter and producing smooth, physically plausible 3D pose sequences.
  • Item
    RAW: Robust AvatarWatermarking - Benchmarking and Baseline
    (The Eurographics Association, 2026) Parry, Jack; Saunders, Jack; Namboodiri, Vinay; Musialski, Przemyslaw; Lim, Isaak
    Digital avatar watermarking presents unique challenges: avatars are routinely post-processed with background replacement, reframing, and format conversion before deployment. We introduce RAW (Robust Avatar Watermarking), a benchmark comprising 50 synthetic avatar videos from 5 commercial providers and 6 attacks simulating real-world avatar workflows. Evaluating 7 existing methods reveals that avatar-specific attacks such as background removal significantly degrade watermark recovery. We propose WALT (Watermarking Avatars with Learned Textures), which embeds watermarks in UV texture space via 3D face reconstruction. WALT achieves the highest robustness to zoom attacks (92.4%) while maintaining strong performance on background removal (95.6%). We release our benchmark to facilitate research into avatar-specific watermarking.
  • Item
    Beyond FID: Human Perceptual Judgments Reveal Systematic Blind Spots in GAN Face Evaluation
    (The Eurographics Association, 2026) Nierula, Birgit; Melnik, Anna; Barthel, Florian; Brama, Aileen; Hilsmann, Anna; Eisert, Peter; Nikulin, Vadim V.; Gaebler, Michael; Klotzsche, Felix; Chen, Yonghao; Stephani, Tilman; Bosse, Sebastian; Musialski, Przemyslaw; Lim, Isaak
    Generative Adversarial Networks (GANs) can synthesize highly realistic facial images from random noise vectors. The Fréchet Inception Distance (FID) is widely used as a standard metric to automatically evaluate the quality of GAN-generated images. However, it remains unclear to what extent this statistical measure reflects human perceptual judgments, which ultimately define image realism in practical applications. To address this, we conducted a psychophysical study in which participants (n = 20) performed a two-alternative forced-choice task, assessing actual photographs and GAN-generated images as real or fake. We show that while FID provides a reliable global ordering of image quality, it systematically fails for localized semantic artifacts (e.g., eyewear and skin texture) that disproportionately affect human realness judgments. This demonstrates that FID and human perception are not merely noisy versions of the same signal, but that FID has systematic blind spots for localized semantic artifacts that disproportionately drive human realism judgments.
  • Item
    RetiDiff: Stable Underwater Image Color Reconstruction Based on Retinex and Diffusion Distillation
    (The Eurographics Association, 2026) Qiu, Wenyao; Zhou, Zhuang; Zhang, Xin; Chen, Jiayi; Zhou, Shiping; Tao, Ran; Musialski, Przemyslaw; Lim, Isaak
    Underwater image color reconstruction remains challenging due to wavelength-dependent light absorption and scattering that cause severe color casts and visibility degradation. We propose RetiDiff, a Retinex-guided diffusion distillation framework that couples a physics-aware diffusion prior with a lightweight Retinex-based UNet for stable, single-pass color restoration. A conditional Diffusion Transformer (DiT), pretrained on physics-guided underwater–terrestrial pairs, is frozen and distilled via Score Distillation Sampling (SDS) into a Retinex-UNet that predicts reflectance R and illumination L. This distillation transfers domain-agnostic color priors while mitigating cross-domain feature entanglement and avoiding iterative diffusion. To further suppress artifacts from imperfect Retinex separation, an Inter-Component Residual (ICR) regularization penalizes cross-component correlation and gradient co-occurrence, reducing halos, ghosting, and color drift while preserving structural fidelity. Extensive experiments on UIEB, LSUI, and TEST-U45 demonstrate state-of-the-art perceptual quality and LAB-space fidelity, with RetiDiff achieving comparable or superior performance to diffusion-based baselines while requiring far fewer parameters, lower FLOPs, and an order-of-magnitude faster inference.
  • Item
    Controllable Cinemagraph Generation from A Still Image
    (The Eurographics Association, 2026) Le, Van Thanh; Ito, Daichi; Mahapatra, Aniruddha; Mai, Long; Singh, Krishna Kumar; Kulkarni, Kuldeep; Liu, Feng; Fu, Yun; Yoon, Jae Shin; Musialski, Przemyslaw; Lim, Isaak
    We present a training-free framework for controllable cinemagraph generation, enabling the creation of hybrid visuals that combine still imagery with subtle, localized motion. Unlike prior approaches, which offer limited spatial and temporal control, our method allows users to explicitly specify static regions through user-provided masks and to modulate motion intensity across different areas of the scene. Built upon text-guided image-to-video diffusion models, we introduce a soft latent blending strategy that leverages the user-specified mask to seamlessly generate foreground motion while preserving a frozen background. In addition, we propose a new temporal spacing representation which is compatible in positional embedding space to enable fine-grained adjustment of motion characteristics—such as speed and amplitude—within a single video. To avoid motion collapse and unnatural dynamics caused by strong constraints on the first and last frames (i.e., enforcing identical frames), we introduce a two-stage generation strategy that first generates unconstrained motion and then softly enforces seamless looping to the initial frame. Our approach produces high-quality, user-controllable cinemagraphs with precise spatial and temporal fidelity, significantly expanding creative flexibility compared to existing methods.
  • Item
    On the Accuracy of Surface Scattering Theories
    (The Eurographics Association, 2026) Avolio, Matthew; D'Eon, Eugene; Steinberg, Shlomi; Musialski, Przemyslaw; Lim, Isaak
    Surface scattering models are typically validated by fitting measured reflectance data. However, predictive rendering requires deriving appearance directly from surface geometry. We test the predictive accuracy of Microfacet and Generalized Harvey-Shack (GHS) theories by comparing renderings derived strictly from measured surface profiles (AFM/profilometry) against photographs of metal samples. We demonstrate that no current model succeeds across all angles; specifically, Microfacet theory fails at grazing angles where optical roughness vanishes. Additionally, we analytically link the Trowbridge-Reitz (GGX) distribution to the K-correlation PSD. This suggests that the popular "GGX look" arises from wave-optical effects on Gaussian (i.e. Beckmann NDF) surfaces, rather than suggesting an underlying non-Gaussian geometry.
  • Item
    On Cosine Prior Distributions for Neural Path Guiding
    (The Eurographics Association, 2026) Gutsch, Jan-Luca; Dereviannykh, Mikhail; Hanika, Johannes; Musialski, Przemyslaw; Lim, Isaak
    Monte Carlo rendering efficiency depends on the quality of importance sampling distributions. Neural path guiding addresses this by learning adaptive distributions using normalizing flows, which transform simple prior distributions into target distributions. We explore how prior distribution choice affects learning efficiency and show that aligning the prior with components of the rendering integral simplifies the learning task, enabling the use of smaller models. Our cosine-distributed prior, matching the cosine-weighted hemisphere term of the rendering equation, achieves faster convergence and lower noise than uniform priors, with particularly strong improvements in scenes with high geometric complexity.
  • Item
    A Delaunay Keyer for Colorspace-Local Matte Extraction and Spill Suppression
    (The Eurographics Association, 2026) Criddle, Isaac; Holladay, Seth; Egbert, Parris; Musialski, Przemyslaw; Lim, Isaak
    Keying is a fundamental, common, and expensive problem in visual effects, consisting of two tasks: matte extraction and screen spill removal. Today’s keyers often fall short on common edge cases including unevenly lit screens, fine edge detail, and sky replacements. Keying thus continues to be a laborious, complex and expensive problem. Our keying algorithm interpolates over a Delaunay tetrahedralization of user-given sample colors to increase flexibility and expressive power. We compare our method to the visual effects industry’s standard keyers, achieving comparable or superior results for matting and spill suppression.
  • Item
    Implementation is Illustration: Zero-Overhead On-Demand Visualization for High-Performance Linear Algebra
    (The Eurographics Association, 2026) Rautek, Peter; Musialski, Przemyslaw; Lim, Isaak
    Complex mathematical algorithms are often difficult to debug and explain. Traditional approaches to visualization usually require writing separate code, which often leads to divergence between the actual implementation and the visualization. We present a prototype implementation of "Implementation is Illustration" (iii), an experimental concept designed to bridge the gap between high-performance mathematical code and visual explanation. It explores a design where developers write standard mathematical code that can double as its own visualization, with zero runtime overhead when visualization is disabled. By augmenting standard types with optional hook capabilities, the library provides a "Visual Mode" for recording operations and a "Fast Mode" that completely compiles away the overhead. This allows developers to rapidly iterate between visualization and high performance execution. We validate this concept through a set of examples and performance benchmarks.
  • Item
    Capsule: Efficient Player Isolation for Datacenters
    (The Eurographics Association, 2026) Du, Zhouheng; Davari, Nima; Li, Li; Loi, Wei Sen; Kodirov, Nodir; Musialski, Przemyslaw; Lim, Isaak
    Cloud gaming is increasingly popular. A challenge for cloud providers is to keep datacenter utilization high: a non-trivial task due to the diversity of hardware and applications. We introduce Capsule, a mechanism to seamlessly share datacenter resources across multiple players. We implemented Capsule in the Open 3D Engine (O3DE). Our evaluations show that Capsule increases datacenter resource utilization by accommodating up to 2.25x more players, without degrading user experience. This is the product of Capsule using up to 1.43x less GPU, 3.11x less VRAM, 3.7x less CPU, and 3.87x less RAM compared to the baseline. Capsule is also application-agnostic, i.e., no changes were required to run applications on the Capsule-based O3DE. Our experiences with four applications, three servers with different hardware specifications, including one with four GPUs, and a multi-server cluster shows that the Capsule design can be adopted by other game engines to increase datacenter utilization.
  • Item
    VisACD: Visibility-Based GPU-Accelerated Approximate Convex Decomposition
    (The Eurographics Association, 2026) Fokin, Egor; Savva, Manolis; Musialski, Przemyslaw; Lim, Isaak
    Physics-based simulation involves trade-offs between performance and accuracy. In collision detection, one trade-off is the granularity of collider geometry. Primitive-based colliders such as bounding boxes are efficient, while using the original mesh is more accurate but often computationally expensive. Approximate Convex Decomposition (ACD) methods strive for a balance of efficiency and accuracy. Prior works can produce high-quality decompositions but require large numbers of convex parts and are sensitive to the orientation of the input mesh. We address these weaknesses with VisACD, a visibility-based, rotation-equivariant, and intersection-free ACD algorithm with GPU acceleration. Our approach produces high-quality decompositions with fewer convex parts, is not sensitive to shape orientation, and is more efficient than prior work.
  • Item
    Tetrahedron-Tetrahedron Intersection and Volume Computation Using Neural Networks
    (The Eurographics Association, 2026) Pedro, Erendiro ; Meißenhelter, Hermann; Zachmann, Gabriel; Musialski, Przemyslaw; Lim, Isaak
    Narrow-phase collision detection is a critical bottleneck in physics-based simulation, traditionally relying on exact but sequential CPU-bound algorithms that struggle to exploit massive GPU parallelism. In this work, we reframe the tetrahedron-tetrahedron intersection as a learned inference problem. We leverage principles of hierarchical permutation invariance—derived from DeepSets and PointNet—to construct a neural architecture that jointly predicts intersection status and volumetric overlap directly from vertex coordinates. We generate a large dataset of 12 million pairs covering five distinct tetrahedron pair configurations, ensuring the model learns robust geometric decision boundaries. Our most accurate model achieves 99.3% mean classification accuracy across five contact configurations and predicts overlap volume with MAE ≈ 2.88×10−3 (for positive volumes) while remaining entirely GPU-resident. On consumer-grade GPU, our pipeline outperforms the exact CGAL library by 93× in detection speed and over 16,821× when volume computation is included. This “volume-for-free” paradigm offers a transformative trade-off for real-time applications where approximate is acceptable but ultra-fast geometric reasoning is paramount.
  • Item
    ConJEB: A Large Elastic Contact Jet Engine Bracket Quadratic Program Dataset
    (The Eurographics Association, 2026) Ferreira, Stephanie; Giebel, Andreas; Mueller-Roemer, Johannes Sebastian; Musialski, Przemyslaw; Lim, Isaak
    Quadratic programs (QPs) are central to many graphics applications, especially large-scale physically based animation, where they lead to sparse problems with hundreds of thousands to millions of degrees of freedom. Yet, benchmark datasets for such large and sparse QPs are vastly missing. We address this by extending the SimJEB dataset with explicit contact handling and by releasing a new collection of QPs derived from these scenarios. Our setup replaces the abstract contact force in the original GE Jet Engine Bracket Challenge with an explicit cylinder mesh pin, yielding realistic large-scale sparse QPs with up to nearly two million degrees of freedom. We describe the dataset construction and characteristics, and provide it to support fair and reproducible evaluation of QP solvers in graphics and related fields.
  • Item
    Computational Design of Forced-Perspective Structures
    (The Eurographics Association, 2026) Watanabe, Sarika; Fukusato, Tsukasa
    This paper presents a computational method for designing forced-perspective structures. Our model focuses on empirical rules that the brain uses to perceive 3D scenes from 2D views and computes the width and height of each unit in man-made repeating objects (e.g., buildings and roads) that can intentionally emphasize or suppress our perspective from a specified viewpoint. Examples of generated structures and the results of a cognitive experiment are shown to demonstrate the flexibility and the effectiveness of our method.
  • Item
    What a ComfortableWorld: Ergonomic Principles Guided Apartment Layout Generation
    (The Eurographics Association, 2026) Nieciecki, Piotr; Plocharski, Aleksander; Musialski, Przemyslaw
    Current data-driven floor plan generation methods often reproduce the ergonomic inefficiencies found in real-world training datasets. To address this, we propose a novel approach that integrates architectural design principles directly into a transformer-based generative process. We formulate differentiable loss functions based on established architectural standards from literature to optimize room adjacency and proximity. By guiding the model with these ergonomic priors during training, our method produces layouts with significantly improved livability metrics. Comparative evaluations show that our approach outperforms baselines in ergonomic compliance while maintaining high structural validity.
  • Item
    Beyond Segmentation: Structurally Informed Facade Parsing from Imperfect Images
    (The Eurographics Association, 2026) Janicki, Maciej; Plocharski, Aleksander; Musialski, Przemyslaw
    Standard object detectors typically treat architectural elements independently, often resulting in facade parsings that lack the structural coherence required for downstream procedural reconstruction. We address this limitation by augmenting the YOLOv8 training objective with a custom lightweight alignment loss. This regularization encourages grid-consistent arrangements of bounding boxes during training, effectively injecting geometric priors without altering the standard inference pipeline. Experiments on the CMP dataset demonstrate that our method successfully improves structural regularity, correcting alignment errors caused by perspective and occlusion while maintaining a controllable trade-off with standard detection accuracy.
  • Item
    Differentiable Objectives for 3D Scene Relighting via Gradient Descent on OLAT Basis Coefficients
    (The Eurographics Association, 2026) Savage, Anson; Egbert, Parris; Holladay, Seth
    Designing effective lighting is an iterative and often time-consuming process. This work contributes to automatic lighting design research by presenting a render-engine-agnostic optimization routine: gradient descent on RGB multipliers of one-light-at-a-time (OLAT) basis images. We propose several objective functions to accomplish lighting tasks and show that our method is capable of quickly and effectively exploring different lighting styles using either text prompts or reference images.
  • Item
    2D-SuGaR: Surface-Aware Gaussian Splatting for Geometrically Accurate Mesh Reconstruction
    (The Eurographics Association, 2026) Gupta, C. R. Prajwal; Sheth, Divyam; Ha, Jinjoo; Ostrek, Mirela; Thies, Justus
    3D Gaussian Splatting enables the reconstruction of a volumetric scene representation from multi-view images that allows for real-time novel-view point synthesis, however, it struggles with recovering an accurate surface geometry. While 2D Gaussian Splatting (2DGS) addresses this through surface-aligned primitives, its performance depends critically on the initialization quality. Reliance on Structure-from-Motion (SfM) limits the initialization flexibility as well. In this work, we present two key contributions to enhance 2DGS and the extraction of a clean surface mesh. First, we incorporate monocular depth and normal priors for robust initialization, coupled with a clustering-based pruning strategy to eliminate degenerate Gaussians. Second, we introduce a joint mesh-Gaussian refinement similar to SuGaR, that relaxes the strict 2D constraint by transitioning to 3D primitives, providing stronger training signals. Evaluated on the DTU dataset, our method achieves state-of-the-art mesh reconstruction with a Chamfer Distance of 0.67, outperforming prior methods.
  • Item
    OT-UVGS: Revisiting UV Mapping for Gaussian Splatting as a Capacity Allocation Problem
    (The Eurographics Association, 2026) Kim, Byunghyun
    UV-parameterized Gaussian Splatting (UVGS) maps an unstructured set of 3D Gaussians to a regular UV tensor, enabling compact storage and explicit control of representation capacity. Existing UVGS, however, uses a deterministic spherical projection to assign Gaussians to UV locations. Because this mapping ignores the global Gaussian distribution, it often leaves many UV slots empty while causing frequent collisions in dense regions. We reinterpret UV mapping as a capacity-allocation problem under a fixed UV budget and propose OT-UVGS, a lightweight, separable one-dimensional optimal-transport-inspired mapping that globally couples assignments while preserving the original UVGS representation. The method is implemented with rank-based sorting, has O(N logN) complexity for N Gaussians, and can be used as a drop-in replacement for spherical UVGS. Across 184 object-centric scenes and the Mip-NeRF dataset, OT-UVGS consistently improves peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and Learned Perceptual Image Patch Similarity (LPIPS) under the same UV resolution and per-slot capacity (K=1). These gains are accompanied by substantially better UV utilization, including higher non-empty slot ratios, fewer collisions, and higher Gaussian retention. Our results show that revisiting the mapping alone can unlock a significant fraction of the latent capacity of UVGS.
  • Item
    VkSplat: High-Performance 3DGS Training in Vulkan Compute
    (The Eurographics Association, 2026) Chen, Jingxiang; Ibrahim, Mohamed; Liu, Yang
    We present VkSplat, a high-performance, cross-vendor 3D Gaussian Splatting (3DGS) training pipeline implemented fully in Vulkan compute, addressing performance and compatibility limitation of existing training pipelines. With various optimizations, we achieve 3.3× speed and 33% VRAM reduction over CUDA+PyTorch baseline, maintaining quality, and demonstrating compatibility across GPU vendors. To the best of our knowledge, this is the first fully-Vulkan-based 3DGS training pipeline that achieves state-of-the-art performance. Code: https://github.com/harry7557558/vksplat
  • Item
    RETA3D: Real-Time Animatable 3D Gaussian Head Generation
    (The Eurographics Association, 2026) Chen, Shu-Yu; Qiu, Chunshuo; Liu, Feng-Lin; Cao, Yanpei; Fu, Hongbo; Gao, Lin
    3D avatar GANs (generative adversarial networks) learn 3D priors from extensive collections of 2D portrait images. However, existing 3D avatar GANs either struggle with real-time performance or lack 3D consistency. To address these issues, we present RETA3D, a novel 3D GAN framework leveraging the efficiency of 3D Gaussian Splatting (3DGS). Our core contribution is a consecutive mesh-binding 3D Gaussian representation that tightly integrates 3D Gaussians with a FLAME mesh template via a novel local coordinate system defined by surface normals and head pose to ensure consistent animation. We also introduce a dynamic texture generation framework that separates static and dynamic texture components, significantly improving reenactment speed. This framework generates a static texture once and efficiently computes dynamic texture updates per-frame using a compact neural network conditioned on FLAME parameters.
  • Item
    Voxel Deformation-Aware Neural Intersection Function
    (The Eurographics Association, 2026) Kao, Chih-Chen; Makowski, Grzegorz; Fujieda, Shin; Harada, Takahiro
    We extend the Locally-Subdivided Neural Intersection Function (LSNIF) to support parameterized deformable and animated geometry. Our approach introduces a rest-space and deformed-space formulation inspired by meshless rendering, allowing ray samples to be mapped back to a canonical space where a single neural network represents geometry consistently across poses without retraining. To maintain accuracy under deformation-aware training, we incorporate scale-invariant distance regression, uncertainty-weighted multi-task learning, and a hybrid positional-grid encoding. The resulting method preserves the compactness and efficiency of LSNIF while enabling robust neural intersection prediction for dynamic geometry.
  • Item
    Robust Ray–Surface Intersections for Algebraic Surfaces
    (The Eurographics Association, 2026) Szente, Péter; Karikó, Csongor Csanád; Valasek, Gábor
    We present a robust method for rendering algebraic surfaces on the GPU using only single-precision arithmetic. While fitting the surface function to a polynomial along the view ray is efficient, it typically suffers from numerical instability even at moderate degrees. We address this by employing error-free transforms to emulate higher precision without the performance cost of standard double-precision types. We show that the resulting polynomial fit can supply data for inferring directional Lipschitz bounds and we propose a new lower and upper bound on Bézier functions. Additionally, we propose a modification to Yuksel’s bracketed Newton method that uses the fitted polynomial solely to isolate monotonous segments, while the final root refinement relies on bisection of the original implicit function. This strategy ensures numerical stability and register efficiency on consumer graphics hardware. We demonstrate our results on rendering various degree algebraic surfaces.
  • Item
    Helper-Lane Optimized Triangulation of Polygons
    (The Eurographics Association, 2026) Bene, Róbert; Valasek, Gábor
    We propose to use total edge length minimizing triangulation of polygons for high-performance rendering. We show that this reduces the number of helper-lane invocations and improves effective GPU utilization. We show that helper-lane count is an indicator of performance, however, it is not the only factor. We propose an edge flipping postprocessing algorithm to improve the rendering performance of arbitrary baseline triangulations. Our comparisons are carried out on vector graphics data.
  • Item
    EUROGRAPHICS 2026: Short Papers Frontmatter
    (The Eurographics Association, 2026) Lim, Isaak; Musialski, Przemyslaw