3D Aesthetics is significant in digital design, shaping how users experience real-time 3D content in games, VR, and product design. However, creating aesthetically pleasing shapes remains challenging due to diverse subjective standards and the lack of tools that support aesthetics-driven editing. Users often rely on intuition without explicit guidance on visual appeal, making aesthetics refinement slow, inconsistent, and cognitively demanding, particularly in fast-paced, iterative workflows. To address this challenge, we conducted in-depth interviews with design experts to identify challenges in aesthetics-oriented modeling workflows. Based on the findings, we developed Aesthetic3D , a 3D modeling interface that provides real-time aesthetics scores learned from human perceptual data. Furthermore, Aesthetic3D seamlessly integrates the learned aesthetics measures into intuitive editing operations, enabling aesthetics-driven exploration and refinement of shape geometry. We evaluated Aesthetic3D through an ablation study, an open-ended study, and three generalization evaluations. Comprehensive experiments show that with Aesthetic3D , users can easily and effectively enhance the aesthetics appeal of 3D shapes.
Implicit surfaces offer distinct advantages over traditional boundary representations, including infinite resolution, low memory footprint, smooth geometry by construction, and support for non-destructive modeling. In this work, we introduce a method for localizing geometric detail in a way that preserves the mathematical properties required for accurate and efficient rendering using sphere tracing. Our contributions include novel procedural modeling techniques that expand the range of repetition patterns achievable in implicit surfaces; an interpolation-based approach that maintains field correctness while remaining computationally efficient; and a cache-based acceleration strategy that significantly improves the rendering performance of domain-repeated implicit geometries.
Welcome to the 2026 I3D PACMCGIT issue. PACMCGIT is a specialized journal in the collection of Proceedings of the ACM (PACM) that is specifically focused on publishing research of the highest quality within the broad domains of computer graphics and interactive techniques. This issue is a collection of the papers selected and presented at the 30th ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games (I3D) 2026. Continuing the tradition, the I3D collection of papers constitutes the first PACMCGIT issue of the year, this time marking the beginning of PACMCGIT’s ninth year presenting the publications of I3D. I3D 2026 received 41 submissions. Following a rigorous double-blind peer-reviewing process with 122 reviews, including a second review cycle by a panel of experts from the I3D 2026 international program committee, 16 papers were accepted to be presented at the conference, resulting in an acceptance rate of \(39\%\) . This year, one paper was desk-rejected as it was clearly incomplete. Paper topics span various exciting areas, from physical simulation and animation to real-time rendering, ray-tracing, VR, and neural rendering. These papers appeared in talk presentations at the ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games (I3D) in May 2026 at Lucasfilm in San Francisco (CA).
The generation of consistent multi-view images is essential for diffusion-based 3D texture synthesis, a key component of modern 3D content pipelines. However, existing methods remain weakly coupled across viewpoints, often exhibiting inconsistencies between views that result in artifacts such as ghosting, misalignment, and texture distortions. We introduce PixelSync, a training-free mechanism that explicitly enforces multi-view consistency during the diffusion process. Our method operates in two stages. First, we compute visibility-aware pixel correspondences between known views in a fast pre-computation step by utilizing geometric information obtained from a given 3D model. Then, during the diffusion process, we enforce consistency through two complementary algorithms: (i) attention synchronization, by modifying the attention scores in the U-Net’s down- and mid-blocks, and (ii) latent synchronization, by progressively aligning features on their corresponding shared pixels. PixelSync can be seamlessly integrated into pre-trained diffusion models without retraining, offering a lightweight and effective solution. We demonstrate the effectiveness of our approach for generative texture synthesis by plugging PixelSync into established multi-view diffusion backbones, showcasing reduced cross-view artifacts and improved global coherence across standard quantitative and qualitative evaluations.
This paper addresses the problem of unpaired shape translation on 3D point clouds. While prior methods typically rely on global latent vectors or spatially structured grids, such representations often lack the flexibility to capture both semantic structures and fine-grained geometric details. To address this, we propose operating directly in a structured token space, where tokens are pretrained through masked autoencoding. Unlike rigid spatial grids that impose fixed layouts, our tokens naturally adapt to geometric variations while maintaining semantic coherence. This structured yet flexible latent space enables semantically meaningful and geometrically precise transformations. A transformer-based translator is proposed to manipulate these tokens. This gated dual-branch translator enables detail-preserving and topology-aware shape translation across categories. Experiments on challenging tasks, such as chair-to-table transformations, demonstrate that our approach outperforms existing methods in preserving both global structure and part-level details.
