arXiv:2605. 01171v2 Announce Type: replace-cross Abstract: Despite recent progress, recovering parametric CAD construction sequences from geometric input, such as meshes or point clouds, is a key challenge for design and manufacturing, as existing CAD reconstruction and generation methods are largely restricted to difficult-to-edit formats like meshes or Breps or editable simple sketch-and-extrude pipelines and low-complexity datasets.
By Ghadi Nehme, Eamon Whalen, Faez Ahmed
Reconstructing an editable CAD model from a 3D shape remains a challenging engineering task. Existing methods can propose CAD operations, but no single source of proposals works equally well across di...
CADFather is an autonomous system that reconstructs editable CAD models from 3D meshes by coordinating vision‑language assistants, learned and algorithmic tools, and numerical optimization. It evaluates and refines candidate CAD programs, maintaining alternatives for each part and preserving the best valid result throughout reconstruction. The system uses pretrained models without additional training and is evaluated on multiple benchmark datasets for reconstruction quality and execution validity.
By Gennadiy Savrasov, Maksim Elistratov, Nikita Gavrilov, Albert Garifullin, Oleg Pavlov, Soslan Kabisov, Vladimir Frolov, Anton Konushin, Andrey Kuznetsov, Dmitrii Zhemchuzhnikov
The paper introduces a generalizable deformation learning framework that reconstructs 3D objects by deforming a category-level shape template to match a monocular observation. It employs a geometry-guided feature modeling mechanism to enrich foundation features with template topology, creating a geometry-aware representation that is explicitly correlated with the target observation for precise deformation. A view-adaptive feature aggregation module further bridges the gap between the fixed template and arbitrary target views by leveraging multi-view template features and camera poses, ensuring robust feature alignment across diverse viewpoints.
By Yiyao Ma, Kai Chen, Zhongxiang Zhou, Zhuheng Song, Dongsheng Xie, Zelong Tan, Rong Xiong, Qi Dou
CADKnitter is a compositional CAD generation framework that uses geometric-guiding cues to steer diffusion sampling, enabling the creation of complementary CAD parts that satisfy both geometric constraints of an existing model and semantic constraints from a text prompt. The authors introduce KnitCAD, a dataset of over 310,000 CAD models paired with textual prompts and assembly metadata to support training and evaluation. Experiments show that CADKnitter outperforms state‑of‑the‑art baselines by a clear margin.
By Tri Le, Khang Nguyen, Baoru Huang, Tung D. Ta, Anh Nguyen
CADReasoner is a model that iteratively refines its CAD reconstruction by comparing the input shape with the predicted mesh, outputting a runnable CadQuery Python program. It fuses multi‑view renders and point clouds and uses a scan‑simulation protocol to bridge realism gaps during training and evaluation. The approach achieves state‑of‑the‑art performance on DeepCAD, Fusion 360, and MCB benchmarks for both clean and scan‑sim tracks.
By Soslan Kabisov, Vsevolod Kirichuk, Andrey Volkov, Gennadii Savrasov, Marina Barannikov, Anton Konushin, Andrey Kuznetsov, Dmitrii Zhemchuzhnikov
arXiv:2603. 05607v2 Announce Type: replace-cross Abstract: Computer-Aided Design (CAD) relies on structured and editable geometric representations, yet existing generative methods are constrained by small annotated datasets with explicit design histories or boundary representation (BRep) labels.
By Mohammad Sadil Khan, Muhammad Usama, Rolandos Alexandros Potamias, Didier Stricker, Muhammad Zeshan Afzal, Jiankang Deng, Ismail Elezi
Mesh subdivision is a fundamental operation for converting coarse, editable meshes into high-resolution surfaces, with broad applications in digital asset creation. Classical rule-based schemes rely on fixed local refinement rules and often produce over-smoothed surfaces.
The paper introduces a real‑world CAD segmentation benchmark focused on fastener‑relevant hole detection, a task that demands human annotation and presents higher geometric complexity than typical synthetic datasets. It presents two rule‑based methods—a graph‑wire‑search and a multi‑view approach—and trains learning‑based models (UV‑Net and AAGNet) on a newly annotated engineering CAD dataset. Extensive experiments evaluate parameter tuning, cross‑domain generalisation, and performance across dataset sizes and IoU thresholds, revealing that rule‑based methods generalise without training data while learning‑based methods excel with sufficient annotations but are less robust to domain shift.
"whyItMatters":"The study provides a realistic benchmark and actionable insights into the trade‑offs between data requirements and robustness for CAD segmentation, informing future research and practical applications in engineering tasks."
By Kai Sch\"utte, Edgar Heinert, Matthias Rottmann
arXiv:2610.33575v1 Announce Type: new
Abstract: Existing methods for CAD reconstruction from 3D scans achieve strong performance on standard benchmarks, yet their effectiveness often degrades in real...
By Ahmet Serdar Karadeniz, Dimitrios Mallis, Parsa Akbari Chianeh, Kseniya Cherenkova, Anis Kacem, Djamila Aouada
arXiv:2605. 10873v2 Announce Type: replace-cross Abstract: Recovering editable CAD programs from images or 3D observations is central to AI-assisted design, but progress is difficult to measure because existing evaluations are fragmented across datasets, modalities, and metrics.
By Anna C. Doris, Jacob Thomas Sony, Ghadi Nehme, Era Syla, Amin Heyrani Nobari, Faez Ahmed
The paper introduces a lightweight RGBD-to-3D refinement technique that improves single-view 3D reconstructions by correcting visible surfaces through bipartite matching with back-projected depth points, mirroring corrections across detected symmetry planes, and propagating them via a smoothness solver. The method is closed-form, significantly faster than optimization-heavy refinements, and works plug-and-play with any RGB-to-3D reconstructor without retraining. Experiments on GSO and OmniObject3D with five backbones show consistent accuracy gains, especially on symmetric objects, and the approach also benefits RGB-D-to-mesh reconstructions and real noisy sensor data.
By Ahyun Seo, Minsu Cho