Towards Stellarator Geometry Optimisation for Nuclear Fusion
We present a local geometry refinement method and a 2D latent representation that took us to the top of the ConStellaration leaderboard on the geometric task in May 2026.
cs.CE↗
arXiv subjects
Florian Bernard
Publications and source records attributed to Florian Bernard.
Unsupervised Pixel-Level Semantic Left-Right Understanding of In-the-Wild Images
While various works address reflective symmetry understanding in 3D data and images, pixel-level semantic left-right prediction of in-the-wild images remains challenging, due to certain difficulties including the lack of 3D information, occlusion, object pose variation, partiality, etc. In this work, we propose an unsupervised learning framework to tackle this challenge. Leveraging recent advances in vertex-wise semantic left-right understanding of 3D data, our unsupervised learning method jointly utilises 3D shape and image datasets to infer pixel-wise semantic left-right predictions in single-view images. In particular, we show that a medium-scale 3D shape dataset comprising mainly of human- and quadruped animal-like shapes, combined with diverse in-the-wild image data, are sufficient to achieve high-quality semantic left-right prediction in images, even for entirely unseen 3D object categories, such as cars or trains. Overall, our approach achieves superior performance in dense pixel-wise semantic left-right predictions on both rendered and in-the-wild image datasets when compared to existing state-of-the-art methods.
cs.CV↗
Hyper-Network Neural Functional Maps for Unsupervised Robust 3D Shape Matching
Functional maps are the cornerstone of recent non-rigid 3D shape matching methods due to their efficiency and performance. However, existing methods struggle with challenging scenarios, such as partiality, topological noise, and raw point clouds. A primary bottleneck is that significant intrinsic distortion prevents truncated spectral bases from being accurately aligned via linear transformations (i.e., functional maps). To address this, we introduce a hyper-network that predicts non-linear neural functional maps (NFM), learned in an unsupervised manner, to better align spectral bases. Specifically, we model the NFM as an MLP with skip-connection to refine standard FM and employ a hyper-network to predict its weights, conditioned on standard FM. Our framework is trained using a novel unsupervised spectral alignment loss. Experiments demonstrate that our approach can be seamlessly integrated into state-of-the-art unsupervised deep functional map pipelines, substantially improving matching accuracy in demanding scenarios.
cs.CV↗
Non-Rigid 3D Shape Correspondences: From Foundations to Open Challenges and Opportunities
Estimating correspondences between deformed shape instances is a long-standing problem in computer graphics; numerous applications, from texture transfer to statistical modelling, rely on recovering an accurate correspondence map. Many methods have thus been proposed to tackle this challenging problem from varying perspectives, depending on the downstream application. This state-of-the-art report is geared towards researchers, practitioners, and students seeking to understand recent trends and advances in the field. We categorise developments into three paradigms: spectral methods based on functional maps, combinatorial formulations that impose discrete constraints, and deformation-based methods that directly recover a global alignment. Each school of thought offers different advantages and disadvantages, which we discuss throughout the report. Meanwhile, we highlight the latest developments in each area and suggest new potential research directions. Finally, we provide an overview of emerging challenges and opportunities in this growing field, including the recent use of vision foundation models for zero-shot correspondence and the particularly challenging task of matching partial shapes.
cs.GR↗
An Integer Linear Programming Approach to Geometrically Consistent Partial-Partial Shape Matching
The task of establishing correspondences between two 3D shapes is a long-standing challenge in computer vision. While numerous studies address full-full and partial-full 3D shape matching, only a limited number of works have explored the partial-partial setting, very likely due to its unique challenges: we must compute accurate correspondences while at the same time find the unknown overlapping region. Nevertheless, partial-partial 3D shape matching reflects the most realistic setting, as in many real-world cases, such as 3D scanning, shapes are only partially observable. In this work, we introduce the first integer linear programming approach specifically designed to address the distinctive challenges of partial-partial shape matching. Our method leverages geometric consistency as a strong prior, enabling both robust estimation of the overlapping region and computation of neighbourhood-preserving correspondences. We empirically demonstrate that our approach achieves high-quality matching results both in terms of matching error and smoothness. Moreover, we show that our method is more scalable than previous formalisms.
cs.CV↗
Symmetry Informative and Agnostic Feature Disentanglement for 3D Shapes
Shape descriptors, i.e., per-vertex features of 3D meshes or point clouds, are fundamental to shape analysis. Historically, various handcrafted geometry-aware descriptors and feature refinement techniques have been proposed. Recently, several studies have initiated a new research direction by leveraging features from image foundation models to create semantics-aware descriptors, demonstrating advantages across tasks like shape matching, editing, and segmentation. Symmetry, another key concept in shape analysis, has also attracted increasing attention. Consequently, constructing symmetry-aware shape descriptors is a natural progression. Although the recent method $\chi$ (Wang et al., 2025) successfully extracted symmetry-informative features from semantic-aware descriptors, its features are only one-dimensional, neglecting other valuable semantic information. Furthermore, the extracted symmetry-informative feature is usually noisy and yields small misclassified patches. To address these gaps, we propose a feature disentanglement approach which is simultaneously symmetry informative and symmetry agnostic. Further, we propose a feature refinement technique to improve the robustness of predicted symmetry informative features. Extensive experiments, including intrinsic symmetry detection, left/right classification, and shape matching, demonstrate the effectiveness of our proposed framework compared to various state-of-the-art methods, both qualitatively and quantitatively.
cs.CV↗
Jailbreaking LLMs Without Gradients or Priors: Effective and Transferable Attacks
As Large Language Models (LLMs) are increasingly deployed in safety-critical domains, rigorously evaluating their robustness against adversarial jailbreaks is essential. However, current safety evaluations often overestimate robustness because existing automated attacks are limited by restrictive assumptions. They typically rely on handcrafted priors or require white-box access for gradient propagation. We challenge these constraints by demonstrating that token-level iterative optimization can succeed without gradients or priors. We introduce RAILS (RAndom Iterative Local Search), a framework that operates solely on model logits. RAILS matches the effectiveness of gradient-based methods through two key innovations: a novel auto-regressive loss that enforces exact prefix matching, and a history-based selection strategy that bridges the gap between the proxy optimization objective and the true attack success rate. Crucially, by eliminating gradient dependency, RAILS enables cross-tokenizer ensemble attacks. This allows for the discovery of shared adversarial patterns that generalize across disjoint vocabularies, significantly enhancing transferability to closed-source systems. Empirically, RAILS achieves near 100% success rates on multiple open-source models and high black-box attack transferability to closed-source systems like GPT and Gemini.
cs.LG↗
Hyper Diffusion Avatars: Dynamic Human Avatar Generation using Network Weight Space Diffusion
Creating human avatars is a highly desirable yet challenging task. Recent advancements in radiance field rendering have achieved unprecedented photorealism and real-time performance for personalized dynamic human avatars. However, these approaches are typically limited to person-specific rendering models trained on multi-view video data for a single individual, limiting their ability to generalize across different identities. On the other hand, generative approaches leveraging prior knowledge from pre-trained 2D diffusion models can produce cartoonish, static human avatars, which are animated through simple skeleton-based articulation. Therefore, the avatars generated by these methods suffer from lower rendering quality compared to person-specific rendering methods and fail to capture pose-dependent deformations such as cloth wrinkles. In this paper, we propose a novel approach that unites the strengths of person-specific rendering and diffusion-based generative modeling to enable dynamic human avatar generation with both high photorealism and realistic pose-dependent deformations. Our method follows a two-stage pipeline: first, we optimize a set of person-specific UNets, with each network representing a dynamic human avatar that captures intricate pose-dependent deformations. In the second stage, we train a hyper diffusion model over the optimized network weights. During inference, our method generates network weights for real-time, controllable rendering of dynamic human avatars. Using a large-scale, cross-identity, multi-view video dataset, we demonstrate that our approach outperforms state-of-the-art human avatar generation methods.
cs.GR↗
Symmetry Understanding of 3D Shapes via Chirality Disentanglement
Chirality information (i.e. information that allows distinguishing left from right) is ubiquitous for various data modes in computer vision, including images, videos, point clouds, and meshes. While chirality has been extensively studied in the image domain, its exploration in shape analysis (such as point clouds and meshes) remains underdeveloped. Although many shape vertex descriptors have shown appealing properties (e.g. robustness to rigid-body transformations), they are often not able to disambiguate between left and right symmetric parts. Considering the ubiquity of chirality information in different shape analysis problems and the lack of chirality-aware features within current shape descriptors, developing a chirality feature extractor becomes necessary and urgent. Based on the recent Diff3F framework, we propose an unsupervised chirality feature extraction pipeline to decorate shape vertices with chirality-aware information, extracted from 2D foundation models. We evaluated the extracted chirality features through quantitative and qualitative experiments across diverse datasets. Results from downstream tasks including left-right disentanglement, shape matching, and part segmentation demonstrate their effectiveness and practical utility. Project page: https://wei-kang-wang.github.io/chirality/
cs.CV↗
4D-CTA Image and geometry dataset for kinematic analysis of abdominal aortic aneurysms
This article presents a dataset used in the article "Kinematics of Abdominal Aortic Aneurysms", published in the Journal of Biomechanics. The dataset is publicly available for download from the Zenodo data repository (https://doi.org/10.5281/zenodo.15477710). The dataset includes time-resolved 3D computed tomography angiography (4D-CTA) images of abdominal aortic aneurysm (AAA) captured throughout the cardiac cycle from ten patients diagnosed with AAA, along with ten patient-specific AAA geometries extracted from these images. Typically, the 4D-CTA dataset for each patient contains ten electrocardiogram (ECG)-gated 3D-CTA image frames acquired over a cardiac cycle, capturing both the systolic and diastolic phases of the AAA configuration. For method verification, the dataset also includes synthetic ground truth data generated from Patient 1's 3D-CTA AAA image in the diastolic phase. The ground truth data includes the patient-specific finite element (FE) biomechanical model and a synthetic systolic 3D-CTA image. The synthetic systolic image was generated by warping Patient 1's diastolic 3D-CTA image using the realistic displacement field obtained from the AAA biomechanical FE model. The images were acquired at Fiona Stanley Hospital in Western Australia and provided to the researchers at the Intelligent Systems for Medicine Laboratory at The University of Western Australia (ISML-UWA), where image-based AAA kinematic analysis was performed. Our dataset enabled the analysis of AAA wall displacement and strain throughout the cardiac cycle using a non-invasive, in vivo, image registration-based approach. The use of widely adopted, open-source file formats (NRRD for images and STL for geometries) facilitates broad applicability and reusability in AAA biomechanics studies that require patient-specific geometry and information about AAA kinematics during cardiac cycle.
cs.CE↗
TwoSquared: 4D Generation from 2D Image Pairs
Despite the astonishing progress in generative AI, 4D dynamic object generation remains an open challenge. With limited high-quality training data and heavy computing requirements, the combination of hallucinating unseen geometry together with unseen movement poses great challenges to generative models. In this work, we propose TwoSquared as a method to obtain a 4D physically plausible sequence starting from only two 2D RGB images corresponding to the beginning and end of the action. Instead of directly solving the 4D generation problem, TwoSquared decomposes the problem into two steps: 1) an image-to-3D module generation based on the existing generative model trained on high-quality 3D assets, and 2) a physically inspired deformation module to predict intermediate movements. To this end, our method does not require templates or object-class-specific prior knowledge and can take in-the-wild images as input. In our experiments, we demonstrate that TwoSquared is capable of producing texture-consistent and geometry-consistent 4D sequences only given 2D images.
cs.CV↗
Fast Globally Optimal and Geometrically Consistent 3D Shape Matching
Geometric consistency, i.e. the preservation of neighbourhoods, is a natural and strong prior in 3D shape matching. Geometrically consistent matchings are crucial for many downstream applications, such as texture transfer or statistical shape modelling. Yet, in practice, geometric consistency is often overlooked, or only achieved under severely limiting assumptions (e.g. a good initialisation). In this work, we propose a novel formalism for computing globally optimal and geometrically consistent matchings between 3D shapes which is scalable in practice. Our key idea is to represent the surface of the source shape as a collection of cyclic paths, which are then consistently matched to the target shape. Mathematically, we construct a hyper product graph (between source and target shape), and then cast 3D shape matching as a minimum-cost circulation flow problem in this hyper graph, which yields global geometrically consistent matchings between both shapes. We empirically show that our formalism is efficiently solvable and that it leads to high-quality results.
cs.GR↗
4Deform: Neural Surface Deformation for Robust Shape Interpolation
Generating realistic intermediate shapes between non-rigidly deformed shapes is a challenging task in computer vision, especially with unstructured data (e.g., point clouds) where temporal consistency across frames is lacking, and topologies are changing. Most interpolation methods are designed for structured data (i.e., meshes) and do not apply to real-world point clouds. In contrast, our approach, 4Deform, leverages neural implicit representation (NIR) to enable free topology changing shape deformation. Unlike previous mesh-based methods that learn vertex-based deformation fields, our method learns a continuous velocity field in Euclidean space. Thus, it is suitable for less structured data such as point clouds. Additionally, our method does not require intermediate-shape supervision during training; instead, we incorporate physical and geometrical constraints to regularize the velocity field. We reconstruct intermediate surfaces using a modified level-set equation, directly linking our NIR with the velocity field. Experiments show that our method significantly outperforms previous NIR approaches across various scenarios (e.g., noisy, partial, topology-changing, non-isometric shapes) and, for the first time, enables new applications like 4D Kinect sequence upsampling and real-world high-resolution mesh deformation.
cs.CV↗
Towards personalised assessment of abdominal aortic aneurysm structural integrity
Abdominal aortic aneurysm (AAA) is a life-threatening condition characterized by the progressive dilation of the aorta, which can lead to rupture if undetected or untreated. Stress-based rupture risk estimation using computational biomechanics has been widely studied; however, it requires wall strength data that cannot be measured in humans in vivo. To overcome this limitation, the goal of this study is to present a new method for biomechanical assessment of AAA via simultaneous consideration of tension and strain in AAA wall. We present a patient-specific, non-invasive method for assessing the structural integrity of the AAA wall using only time-resolved 3D computed tomography angiography (4D-CTA) images and blood pressure data. The proposed approach integrates wall strain (throughout the cardiac cycle) and wall tension analysis to compute a novel index, the Relative Structural Integrity Index (RSII), which quantifies local wall stiffness independently of wall thickness, wall material properties, and blood pressure measurement conditions. We applied our method to twenty patients from three different hospitals to extract visual RSII maps over the AAA wall of each individual patient and to compare the RSII values between aneurysmal and non-aneurysmal aortas in one patient. Our results primarily show similar RSII values across all patients, indicating the consistency of the method. Additionally, we observed patterns consistent with experimental findings reported in the literature: AAA walls exhibited higher stiffness than healthy aortic walls, while localized low-stiffness zones in the AAA wall were predominantly found in the most dilated regions.
cs.CE↗
Towards Full Automation of Geometry Extraction for Biomechanical Analysis of Abdominal Aortic Aneurysm; Neural Network-Based versus Classical Methodologies
Background: For the clinical adoption of stress-based rupture risk estimation in abdominal aortic aneurysms (AAAs), a fully automated pipeline, from clinical imaging to biomechanical stress computation, is essential. To this end, we investigated the impact of AI-based image segmentation methods on stress computation results in the walls of AAAs. We compared wall stress distributions and magnitudes calculated from geometry models obtained from classical semi-automated segmentation versus automated neural network-based segmentation. Method: 16 different AAA contrast-enhanced computed tomography (CT) images were semi-automatically segmented by an analyst, taking between 15 and 40 minutes of human effort per patient, depending on image quality. The same images were automatically segmented using PRAEVAorta2 commercial software by NUREA (https://www.nurea-soft.com/), developed based on artificial intelligence (AI) algorithms, and automatically post-processed with an in-house MATLAB code, requiring only 1-2 minutes of computer time per patient. Aneurysm wall stress calculations were automatically performed using the BioPARR software (https://bioparr.mech.uwa.edu.au/). Results: Compared to the classical semi-automated segmentation, the automatic neural network-based segmentation leads to equivalent stress distributions, and slightly higher peak and 99th percentile maximum principal stress values. However, our statistical analysis indicated that the differences in AAA wall stress obtained using the two segmentation methods are not statistically significant and fall well within the typical range of inter-analyst and intra-analyst variability. Conclusions: Our findings are a steppingstone toward a fully automated pipeline for biomechanical analysis of AAAs, starting with CT scans and concluding with wall stress assessment.
cs.CE↗
Implicit Neural Surface Deformation with Explicit Velocity Fields
In this work, we introduce the first unsupervised method that simultaneously predicts time-varying neural implicit surfaces and deformations between pairs of point clouds. We propose to model the point movement using an explicit velocity field and directly deform a time-varying implicit field using the modified level-set equation. This equation utilizes an iso-surface evolution with Eikonal constraints in a compact formulation, ensuring the integrity of the signed distance field. By applying a smooth, volume-preserving constraint to the velocity field, our method successfully recovers physically plausible intermediate shapes. Our method is able to handle both rigid and non-rigid deformations without any intermediate shape supervision. Our experimental results demonstrate that our method significantly outperforms existing works, delivering superior results in both quality and efficiency.
cs.CV↗
DiscoMatch: Fast Discrete Optimisation for Geometrically Consistent 3D Shape Matching
In this work we propose to combine the advantages of learningbased and combinatorial formalisms for 3D shape matching. While learningbased methods lead to state-of-the-art matching performance, they do not ensure geometric consistency, so that obtained matchings are locally non-smooth. On the contrary, axiomatic, optimisation-based methods allow to take geometric consistency into account by explicitly constraining the space of valid matchings. However, existing axiomatic formalisms do not scale to practically relevant problem sizes, and require user input for the initialisation of non-convex optimisation problems. We work towards closing this gap by proposing a novel combinatorial solver that combines a unique set of favourable properties: our approach (i) is initialisation free, (ii) is massively parallelisable and powered by a quasi-Newton method, (iii) provides optimality gaps, and (iv) delivers improved matching quality with decreased runtime and globally optimal results for many instances.
cs.CV↗
Beyond Complete Shapes: A Benchmark for Quantitative Evaluation of 3D Shape Surface Matching Algorithms
Finding correspondences between 3D deformable shapes is an important and long-standing problem in geometry processing, computer vision, graphics, and beyond. While various shape matching datasets exist, they are mostly static or limited in size, restricting their adaptation to different problem settings, including both full and partial shape matching. In particular the existing partial shape matching datasets are small (fewer than 100 shapes) and thus unsuitable for data-hungry machine learning approaches. Moreover, the type of partiality present in existing datasets is often artificial and far from realistic. To address these limitations, we introduce a generic and flexible framework for the procedural generation of challenging full and partial shape matching datasets. Our framework allows the propagation of custom annotations across shapes, making it useful for various applications. By utilising our framework and manually creating cross-dataset correspondences between seven existing (complete geometry) shape matching datasets, we propose a new large benchmark BeCoS with a total of 2543 shapes. Based on this, we offer several challenging benchmark settings, covering both full and partial matching, for which we evaluate respective state-of-the-art methods as baselines.
cs.CV↗