Skeleton-based implicit surfaces, and convolution surfaces in particular, offer an intuitive and expressive representation for geometric modeling. However, their evaluation and rendering remain computationally demanding, and are often affected by aliasing artifacts. In this paper, we introduce a continuous level of detail framework for the automatic simplification of SCALIS convolution surfaces, achieving up to an order-of-magnitude reduction in rendering time while mitigating aliasing artifacts. Our solution dynamically adapts the blending behavior between shape components to preserve the perceived visual appearance of the surface, while progressively simplifying the underlying skeletal geometry. In contrast to traditional simplification techniques, our method allows controlled changes in topological genus when visually beneficial. Our results demonstrate that the proposed framework effectively reduces visual artifacts and computational cost, while maintaining high fidelity to the overall shape.

This paper presents a novel method, called Deformable Polygonal Flow Matching (DPFM), for the generation of polygonal arrangements such as jigsaw puzzles and floor plans. DPFM is a Flow Matching framework that enables the generation process to deform, rotate, and translate polygons while decoupling these transformations, allowing to toggle them individually. Able to combine the spatial reasoning capabilities of arrangement models with the flexibility of position-based models, it covers a wide range of applications within a unified formulation, from noiseless puzzle solving using rigid alignments to unconstrained floor plan generation.We represent data using a hierarchical graph composed of a topological subgraph encoding connectivity information and semantics (such as room types for floor plans), and a geometrical subgraph encoding the 1D polygonal loop of each shape. DPFM also leverages Flow Matching's arbitrary prior distributions for geometric constraints by designing priors with domain knowledge. Rather than starting the generation process from uninformed distributions, the generation is constrained through the informed priors at the initialization stage. The qualitative and quantitative evaluations of our method, ran on the RPLAN and jigsaw puzzle datasets, demonstrate strong performance. DPFM outperforms task-specific methods, becoming the new state-of-the-art for 2D arrangement generation. Our results show that DPFM is able to solve novel tasks, such as puzzle denoising, where pieces are reconstructed from noisy versions and arranged into a valid puzzle in parallel.

Abstract We present Sketch2Growth , an interactive system for generating a family of 3D branching shapes and animating their growth from a single user‐drawn style sketch. The sketch, seen as a 2D idealized representation of the desired shape, is converted into a concise format called the Seed . This representation consists of a Directed Acyclic Graph (DAG) that encodes the recursive branching topology, and a series of Gaussian Mixture Models (GMMs) that capture statistical visual‐style at branching points and in terms of branch curvature. During synthesis, the Seed serves as support for a lightweight, stochastic generator: Branches are recursively expanded by unfolding the DAG and sampling from learned distributions, which maintains the expected correlation between parent and child branches. Further growth and variability of the generated shapes are achieved through similarity‐based looping in the Seed structure (re‐starting from a similar node), stochastic mutations and clamping. Only requiring a light learning process through the GMMs, Sketch2Growth ensures perceptual similarity between the generated shapes and the input style sketch, as validated through a user study. In addition, interactive handles enable users to fake the effect of external forces during growth. As results show, our system allows to easily model and animate the continuous growth of complex branching shapes inspired by trees, plants, corals or anatomical structures.

Combining physically-based simulation with stylized visual effects not only requires changing the appearance of surfaces, but their shapes as well. We describe a new expressive rendering method for liquid animations, which can be used on top of any preexisting particle-based simulation. Our solution builds on visual particles that carry both water and air distributions, both evolving through particle history based on kinematic information from the simulation. These density fields are combined at each frame to create the implicit iso-surface of interest, rendered in an adapted style. By defining a series of visual particle states, we parametrize this model to capture the typical stylized geometry of water bodies used to highlight dynamic motion in paintings and cartoons, such as elongating droplets, concavities carved at the crest of breaking waves, and stylized air–water mixtures such as bubbles and foam , which we further enhance in 3D scenes using a dedicated stylized surface-color pattern. Regardless of the 2D or 3D nature of the input simulation, our solution maintains temporal coherence and ensures that water bodies keep an approximately constant surface in 2D, resp. volume in 3D, over time. Finally, we conducted a user study to show the effectiveness of our method against state-of-the-art AI-based tools and a hands-on evaluation with a professional artist, in a variety of animation scenarios where stylized shapes are needed. This is an extension of our previous work presented at STAG 2025. It introduces new material, including a refined blending operator, a stylized surface-color pattern, performance metrics, and artist evaluations.

To support the design and subsequent generation of terrestrial planets for use in the creative media, we propose a solution that employs a generative model trained on satellite data from planetary bodies with a defined solid surface, such as the Earth and Mars. A user sketches coarse elevation, landcover, temperature, and precipitation directly onto a globe. Our model then infers high-resolution heightmap and surface appearance layers at planetary scales, with sufficient detail to enable animated flyovers within the exosphere at a distance of a few thousand kilometers from the planet surface. We address the issue of distortion in the mapping from atlas to globe using a quadsphere representation, and the consistency of large-scale geomorphological features by extracting a global river network from the sketch inputs and providing this as conditioning to the diffusion. As our results demonstrate, our generative model provides a balance between: authoring control through a multi-layer painting interface with a satellite image pre-visualization; computation times proportional to the surface area being generated; landscape diversity, displaying, without repetition artefacts, the full range of elevation and landcover features drawn from multiple source planets, and geomorphological plausibility through the provision of a consistent uninterrupted exorheic global river network, where the input sketches allow.

Coarse binary voxel grids (under 1003) provide a simple interface enabling non-expert users to create a coarse approximation of diverse geometric content. Converting voxelized content into piecewise-smooth geometric models that reflect user intent can greatly increase the attractiveness of such interfaces. While multiple methods exist for surfacing binary voxel grids, they by and large target much higher grid resolutions. Applying these to coarse inputs often produces unintuitive results. We introduce VOX2Surf, a novel method for reconstructing user-intended surfaces from coarse binary voxel grids. We observe that a key challenge in achieving this goal is to correctly identify viewer-expected sharp features in these inputs. While human observers easily mentally separate sharp grid edges that are an artefact of the voxel representation from those depicting intended sharp features, existing techniques struggle to distinguish between them. We employ a learning-based approach, targeted at coarse data, to accurately recover the intended sharp features and utilize them for piecewise-smooth surface fitting. After identifying voxels containing sharp features, we employ a novel geometric reconstruction method to extract a curve network from these voxels. We use the loops of this network as the boundaries of our surface patches and use physically based simulation to smooth both the network curves and the surface patches. Extensive comparisons demonstrate that VOX2Surf achieves better approximation of the input voxelized surfaces compared to alternatives. More importantly, our user study confirms that our results are visually significantly better aligned with viewer expectations when presented with the input surfaces than those produced by alternative approaches.

We introduce Arenite, a novel physics-based approach for modeling sandstone structures. The key insight of our work is that simulating a combination of stress and multi-factor erosion enables the generation of a wide variety of sandstone structures observed in nature. We isolate the key shape-forming phenomena: multi-physics fabric interlocking, wind and fluvial erosion, and particle-based deposition processes. Complex 3D structures such as arches, alcoves, hoodoos, or buttes can be achieved by creating simple 3D structures with user-painted erodable areas and vegetation and running the simulation. We demonstrate the algorithm on a wide variety of structures, and our GPU-based implementation achieves the simulation in less than 5 minutes on a desktop computer for our most complex example.

Although intuitive, sketching a closed 3D shape directly in an immersive environment results in an unordered set of arbitrary strokes, which can be difficult to assemble into a closed surface. We tackle this challenge by introducing VRSurf, a surfacing method inspired by a balloon inflation metaphor: Seeded in the sparse scaffold formed by the strokes, a smooth, closed surface is inflated to progressively interpolate the input strokes, sampled into lists of points. These are treated in a divide-and-conquer manner, which allows for automatically triggering some additional balloon inflation followed byfusion ifthe current inflation stops due to a detected concavity. While the input strokes are intended to belong to the same smooth 3D shape, our method is robust to coarse VR input and does not require strokes to be aligned. We simply avoid intersecting strokes that might give an inconsistent surface position due to the roughness of the VR drawing. Moreover, no additional topological information is required, and all the user needs to do is specify the initial seeding location for the first balloon. The results show that VRsurf can efficiently generate smooth surfaces that interpolate sparse sets of unoriented strokes. Validation includes a side-by-side comparison with other reconstruction methods on the same input VR sketch. We also check that our solution matches the user's intent by applying it to strokes that were sketched on an existing 3D shape and comparing what we get to the original one.

Our goal is to generate realistic human motion from natural language. Modern methods often face a trade-off between model expressiveness and text-to-motion alignment. Some align text and motion latent spaces but sacrifice expressiveness; others rely on diffusion models producing impressive motions, but lacking semantic meaning in their latent space. This may compromise realism, diversity, and applicability. Here, we address this by combining latent diffusion with a realignment mechanism, producing a novel, semantically structured space that encodes the semantics of language. Leveraging this capability, we introduce the task of textual motion inversion to capture novel motion concepts from a few examples. For motion synthesis, we evaluate LEAD on HumanML3D and KIT-ML and show comparable performance to the state-of-the-art in terms of realism, diversity, and text-motion consistency. Our qualitative analysis and user study reveal that our synthesized motions are sharper, more human-like and comply better with the text compared to modern methods. For motion textual inversion, our method demonstrates improved capacity in capturing out-of-distribution characteristics in comparison to traditional VAEs.

Preserving semantics, in particular in terms of contacts, is a key challenge when retargeting motion between characters of different morphologies. Our solution relies on a low-dimensional embedding of the character's mesh, based on rigged key vertices that are automatically transferred from the source to the target. Motion descriptors are extracted from the trajectories of these key vertices, providing an embedding that contains combined semantic information about both shape and pose. A novel, adaptive algorithm is then used to automatically select and weight the most relevant features over time, enabling us to efficiently optimize the target motion until it conforms to these constraints, so as to preserve the semantics of the source motion. Our solution allows extensions to several novel use-cases where morphology and mesh contacts were previously overlooked, such as multi-character retargeting and motion transfer on uneven terrains. As our results show, our method is able to achieve real-time retargeting onto a wide variety of characters. Extensive experiments and comparison with state-of-the-art methods using several relevant metrics demonstrate improved results, both in terms of motion smoothness and contact accuracy.

In this work, we investigate whether artistic hatching, popular in pen-and-ink sketches, can be consistently perceived as a depth cue. We illustrate our results by presenting PerceptualLift, a modeling system that exploits hatching to create curved 3D shapes from a single sketch. We first describe a perceptual user study conducted across a diverse group of participants, which confirms the relevance of hatches as consistent clues for inferring curvature in the depth direction from a sketch. It enables us to extract geometrical rules that link 2D hatch characteristics, such as their direction, frequency, and magnitude, to the changes of depth in the depicted 3D shape. Built on these rules, we introduce PerceptualLift, a flexible tool to model 3D organic shapes by simply hatching over 2D hand-drawn contour sketches.

We describe a new rendering technique for expressive liquid surface 2D animation that can be used on top of existing particle-based simulations. We introduce a hybrid particle model that carries both water and air density distribution and can evolve through particle history. These material quantities combined with the kinematics information are then used to generate a scalar field, which can be parameterized to create an implicit iso-surface capturing stylized geometry commonly seen in paintings and cartoons. We propose, in particular, to represent behavior highlighting the dynamical aspect of the scene, such as elongated droplet behavior and curl-like shapes found in breaking waves.

We propose a simple method for the intuitive authoring and efficient animation of virtual tornadoes. Users control the tornado kinematics by sketching two types of curves to specify the initial geometry of the tornado’s core and the profile of the surrounding swirling air, known as the funnel. The first input, a 3D curve, initializes the core as a vortex filament. This filament induces a swirl flow and advects according to its initial curvature, resulting in progressive bending and twisting. The second input consists of one or multiple 2D profile curves that parameterize the Stokes stream function, governing the radial and axial motion of the air around the core and thereby dictate the funnel shape over time. The core and funnel profile are coupled in local frames through closed-form velocities, which together describe the rotation, sliding and uplift within the tornado’s air volume. As shown in our case studies, our method provides a controllable and efficient way to animate visually plausible tornadoes capable of tearing off infrastructure and transporting debris, as well as interacting with uneven terrain.

We propose an automatic method for generating flight summaries of prescribed duration, given any planed 3D trajectory of a flying object. The challenge is to select relevant time-ellipses, while keeping and adequately framing the most interesting parts of the trajectory, and enforcing cinematographic rules between the selected shots. Our solution optimizes the visual quality of the output video both in terms of camera view and film editing choices, thanks to a new optimization technique, designed to jointly optimize the selection of the interesting parts of a flight, and the camera animation parameters over time. To our best knowledge, this solution is the first one to address camera control, film editing, and trajectory summarizing at once. Ablation studies demonstrate the visual quality of the flights summaries we generate compared to alternative methods.

Explosive volcanic eruptions rank among the most terrifying natural phenomena, and are thus frequently depicted in films, games, and other media, usually with a bespoke once-off solution. In this paper, we introduce the first general-purpose model for bi-directional interaction between the atmosphere and a volcano plume. In line with recent interactive volcano models, we approximate the plume dynamics with Lagrangian disks and spheres and the atmosphere with sparse layers of 2D Eulerian grids, enabling us to focus on the transfer of physical quantities such as temperature, ash, moisture, and wind velocity between these sub-models. We subsequently generate volumetric animations by noise-based procedural upsampling keyed to aspects of advection, convection, moisture, and ash content to generate a fully-realized volcanic skyscape. Our model captures most of the visually salient features emerging from volcano-sky interaction, such as windswept plumes, enmeshed cap, bell and skirt clouds, shockwave effects, ash rain, and sheathes of lightning visible in the dark.

Human and animal presence in natural landscapes is initially revealed by the immediate impact of their locomotion, from footprints to crushed grass. In this work, we present an approach to model the effects of virtual characters on natural terrains, focusing on the impact of human locomotion. We introduce a lightweight solution to compute accurate foot placement on uneven ground and infer dynamic foot pressure from kinematic animation data and the mass of the character. A ground and vegetation model enables us to effectively simulate the local impact of locomotion on soft soils and plants over time, resulting in the formation of visible paths. As our results show, we can parameterize various soil materials and vegetation types validated with real-world data. Our method can be used to significantly increase the realism of populated natural landscapes and the sense of presence in virtual applications and games.

Having simple methods of illustration is essential to scientific thinking. To complement the abstract sketches regularly used in cell biology, we propose DynBioSketch, an easy-to-use digital modeling and animation tool, enabling biologists to resort to less simplified representations when necessary without having to call professional artists. DynBioSketch is an interactive sketching system dedicated to the design and communication of biological phenomena at the cellular scale that can be illustrated in a few minutes of animation. Our model integrates 3D modeling, pattern-based design of 3D shape distributions, and sketch-based animation. These elements can be combined to create complex scenarios such as the infection phenomenon on which we focus, allowing a narrative design adapted to communication between researchers or educational applications in biology. Our results, along with a user study conducted with biology researchers, highlight the potential of DynBioSketch in enabling the direct design of dynamic visual summaries that convey relevant information, as shown in our infection case study. By bridging the gap between abstract representations used by experts and more illustrative depictions, DynBioSketch opens a new avenue for communicating biological concepts.