Pith. sign in

REVIEW 3 major objections 6 minor 64 references

CAD-Recode: Reverse Engineering CAD Code from Point Clouds

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read CAD-Recode claims that CAD reverse engineering can be recast as a code-generation problem: a point cloud is translated into executable Python code (using the CadQuery library) that reconstructs the CAD model.

desk verdict Solid new idea, honest ablations, but borrowed baselines and metric-aware reranking inflate the margin; still deserves a serious referee. read the letter →

arxiv 2412.14042 v2 pith:P227LUHT submitted 2024-12-18 cs.CV

Danila Rukhovich , Elona Dupont , Dimitrios Mallis , Kseniya Cherenkova , Anis Kacem , Djamila Aouada This is my paper · ORCID
classification cs.CV
keywords CADreverseengineeringpointcloudslargelanguagemodelsQueryPythoncodegenerationproceduraldatasetsketch-extrudesequencesparametric
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

CAD-Recode claims that CAD reverse engineering can be recast as a code-generation problem: a point cloud is translated into executable Python code (using the CadQuery library) that reconstructs the CAD model. The paper argues this works because pre-trained large language models already know Python syntax, so a relatively small LLM (Qwen2-1.5B) can be fine-tuned with a lightweight point-cloud projector to produce valid CAD programs. To overcome the scarcity of human CAD sequences, the authors procedurally generate one million sketch-extrude models as CadQuery code. On the DeepCAD, Fusion360, and real-world CC3D datasets, the method reports substantially lower Chamfer distance and higher IoU than prior sketch-extrude reverse-engineering methods, and the generated code is directly interpretable by off-the-shelf LLMs for question answering and editing. If these results hold, they imply that LLMs plus a large procedural training set can make CAD reverse engineering practical, producing models that are both accurate and editable.

What carries the argument

The mechanism is a two-part architecture: (1) a point-cloud projector that downsamples the input to 256 points, applies Fourier positional encoding to the coordinates, and linearly projects them into a token sequence; and (2) a fine-tuned Qwen2-1.5B LLM that autoregressively generates CadQuery Python code from those tokens. The third load-bearing element is the procedural dataset generator, which builds 1M valid sketch-extrude sequences via randomized heuristics (sketch primitives, boolean operations, planes, extrusions, then union), verifies each with syntactic and geometric checks, and quantizes parameters to a fixed grid. Together these let the model sidestep custom CAD-language design and instead reuse the LLM's pre-existing code knowledge.

What would settle it

Run a head-to-head comparison in which CAD-SIGNet (and other baselines) are given the same ten-candidate, Chamfer-reranked inference and identical input sampling, and see whether the DeepCAD/Fusion360/CC3D margins persist. Also probe the generalization boundary: feed CAD-Recode point clouds of models containing revolutions, fillets, or B-splines — features absent from its procedural training set — and measure how often the predicted code fails to reconstruct the shape.

Watch

Extended reading notes

Core claim

The central claim is that representing CAD sketch-extrude sequences as CadQuery Python code lets a pre-trained LLM serve as the decoder for point-cloud-to-CAD reconstruction, provided it is trained on a sufficiently large and controllable procedural dataset. The paper reports that CAD-Recode, trained on 1M generated models, achieves a mean Chamfer distance of 0.30 on the DeepCAD test set and 0.35 on Fusion360, versus 3.43 and 7.37 for the previous best method CAD-SIGNet, with IoU rising from 77.6% to 92.0% and from 65.6% to 87.8% respectively; on the scanned CC3D set, median CD drops from 2.90 to 0.31. The authors also show the predicted code can be fed to a general-purpose LLM to answer CAD-specific geometry questions (76.5% accuracy on SGP-Bench, versus 63.2% with CAD-SIGNet's output) and to generate interactive editing sliders. The underlying assertion is that the code representation absorbs most of the 'design intent' burden, so the network only needs to learn a geometry-to-token mapping rather than a new CAD-specific syntax.

Load-bearing premise

The reported gains rely on the evaluation protocol being fair: CAD-Recode uses test-time reranking over ten generated code candidates chosen by Chamfer distance to the input, while the comparison numbers for other methods are taken from earlier papers that may have used different point-cloud sampling or no equivalent reranking.

Editorial extensions

If this is right

  • CAD reverse engineering becomes a language-generation task, so future gains in LLM code ability should transfer directly to reconstruction quality.
  • The procedural dataset generator can be extended to more operations (revolution, fillet, patterns), potentially scaling to more complex real-world models.
  • The code output enables a new workflow: a designer scans an object, gets an editable CadQuery script, and modifies parameters through sliders or natural language.
  • Because the output is executable code, reconstruction validity can be checked by simply running it, which provides a clean training signal for geometric correctness.
  • The ten-candidate reranking suggests a general verification-based inference strategy that other CAD-sequence models could adopt.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the LLM's Python prior is the main source of the gains, then scaling to a larger LLM or a code-specialised model should improve geometry fidelity further; this is a direct testable extension of the paper's stated scaling plans.
  • The procedural dataset distribution is fully controllable, so it could be re-weighted to target specific failure modes (thin features, non-extrusion operations) and thereby reduce the invalid-prediction cases the authors document.
  • The Chamfer-reranking step effectively turns the generative model into a search over a latent code space; a learned verifier could do the same job at a fraction of the compute, and might also detect invalid code before execution.
  • The approach's reliance on CadQuery ties it to that library's kernel; porting the representation to another programmatic CAD API would be a quick way to test how much of the benefit comes from the code format itself versus the LLM's prior.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper proposes CAD-Recode, a method that maps point clouds to executable Python/CadQuery code representing sketch-extrude CAD sequences. The architecture combines a lightweight point-cloud projector (furthest point sampling, Fourier positional encoding, and a linear projection) with a fine-tuned Qwen2-1.5B LLM decoder. The authors introduce a procedurally generated 1M-sample training dataset and evaluate on DeepCAD, Fusion360, and the real-world CC3D dataset, reporting large improvements in Chamfer distance, IoU, and invalidity ratio over prior reverse-engineering methods. They also demonstrate that the generated code can be interpreted by an off-the-shelf LLM for CAD question answering and interactive editing.

Significance. If confirmed, the paper's contributions are significant: a CAD-code representation that is directly executable and LLM-readable, a demonstration that procedural training data can outperform a hand-collected dataset of the same size, and an order-of-magnitude improvement on point-cloud CAD reverse engineering. The no-sampling ablation in Table 3 is a valuable internal control and supports the qualitative conclusion that the method works even without Chamfer-based reranking. The paper also provides detailed dataset-generation algorithms, validation steps (syntax checks, BRepCheck, duplicate detection), and a transparent single-stage training setup, which are strengths. However, the reported magnitude of the improvement is currently hard to verify because the main comparisons mix borrowed baseline numbers with a test-time reranking protocol that optimizes the evaluation metric, and because all results are single-run point estimates.

major comments (3)
  1. [Sections 4.3 and 5.1, Table 1] The test-time procedure of Section 4.3 generates ten code candidates from different point-cloud samplings and selects the one with minimum Chamfer distance to the input cloud. Since this selection criterion is the same metric on which the method is then evaluated, the reported CD values are minima over ten draws, rather than single-sample expectations. Most baseline numbers in Table 1 are borrowed from earlier papers ([34], [21], [12]) and were not produced under this protocol. The no-sampling rows in Table 3 show that the qualitative conclusion is robust for CAD-SIGNet, but the headline 'ten-fold mean CD improvement' and the IoU margins in Table 1 may still be materially inflated by the protocol asymmetry. Please re-run all baselines under a common evaluation harness with identical point-cloud sampling, identical 8192-point CD computation, and identical ten-candidate Chamfer selection, or state the headline comparisons in the single-sample setting and report both.
  2. [Section 5.1, metric definitions] The paper does not specify whether CD and IoU are averaged over all ten generated candidates, over the single selected candidate, or only over executable/valid candidates. This matters because invalid predictions exist (Figure 14) and because Table 3 shows test-time sampling changes the invalidity ratio from 4.9% to 0.4% for the 1M model. If the reported CD/IoU are computed only after discarding invalid codes (or only on the selected candidate), the comparison with baselines that report over all predictions is not apples-to-apples. Please state the exact aggregation rule and, ideally, report both per-candidate and selected-candidate numbers.
  3. [Section 5.1, Tables 1-4 and Table 6] All quantitative results are single-run point estimates with no error bars, confidence intervals, or multiple seeds. The paper repeatedly claims that CAD-Recode 'significantly outperforms' prior methods and that the procedural dataset is beneficial; these claims are supported only by point estimates. Please add multiple-seed runs (or equivalent variance information) for the main comparisons and the key ablations, and perform a significance test or state the number of seeds used.
minor comments (6)
  1. [Appendix F.2] The reference to 'Table 6 of the main paper' should be Table 5, which contains the CAD-QA results.
  2. [Appendix B, Algorithm 2] Generate2DSketch() returns the boundary components of a single generated sketch, yet the loop 'for sketch in sketches' implies multiple sketches; please clarify the pseudocode to make the generation procedure unambiguous.
  3. [Section 5.1, Table 4] Table 4 reports median CD but the caption only says 'CD'; the appendix Table 6 clarifies this, but the main table should state 'median CD' explicitly.
  4. [Appendix D, Table 7 discussion] The sentence that CAD-Recode achieves 'comparable performance to the state-of-the-art on the command type accuracy' understates the actual numbers, since CAD-Diffuser attains 88.5% versus 83.9% for CAD-Recode; please rephrase to be precise about the comparison.
  5. [Figure 17 caption] The caption contains a typo: 'CADSIGNet' should be 'CAD-SIGNet'.
  6. [Abstract and Section 3.2] The paper states that the 1M dataset will be made publicly accessible but does not provide a repository link or a release timeline; adding this information would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CAD-Recode's central claims are supported by external benchmarks and ablations; self-citations are non-load-bearing.

full rationale

CAD-Recode is an empirical systems paper without a formal derivation chain whose conclusion could reduce to its own premises. The core claim—translating point clouds into executable Python CadQuery code—is trained with a next-token NLL objective on a procedurally generated dataset and then evaluated on external public benchmarks (DeepCAD, Fusion360, CC3D). The test-time reranking in Section 4.3 selects one of ten generated codes by minimizing Chamfer distance to the input point cloud; this is an inference-time selection procedure, not a fitted parameter renamed as a prediction, and the paper transparently ablates it. Table 3 shows that without test-time sampling, CAD-Recode still outperforms CAD-SIGNet on DeepCAD (0.75 vs. 6.81 mean CD), so the qualitative conclusion does not depend on the reranking. Borrowed baseline numbers and possible protocol differences are evaluation-fairness concerns, not circularity, because the baseline methods are published systems used for comparison and CAD-Recode is measured on the same public test sets. Self-citations to CAD-SIGNet, TransCAD, and CC3D are used as baselines or dataset sources, not as load-bearing justification for the method's validity. No step satisfies the required test of exhibiting a specific reduction of a claimed result to its own inputs by construction.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The method introduces no new physical or theoretical entities. It relies on an existing library (CadQuery), a pre-trained LLM, and a procedural dataset generator. The free parameters are training and inference choices, not fitted physical constants.

free parameters (5)
  • Downsampled point count n_p = 256
    Input point clouds are reduced to 256 points; ablations at 64, 128, and 256 show sensitivity.
  • Number of test-time candidates = 10
    At inference, 10 code candidates are sampled and reranked by Chamfer distance to the input cloud.
  • Gaussian noise augmentation = std=0.01, probability=0.5
    Applied to input point coordinates during training.
  • LLM training hyperparameters = lr=2e-4, batch=18, 100k iterations
    Chosen by standard practice; no sensitivity analysis is reported.
  • Quantization resolution in dataset generation = 1 unit within [-100,100]
    Limits the smallest representable feature size and explains several invalid predictions.
assumptions (5)
  • domain assumption Point clouds are sampled from meshes of CAD models and contain enough information to infer sketch-extrude sequences.
    The problem formulation in Section 4.1 assumes a ground-truth B-Rep exists and the point cloud approximates it.
  • domain assumption All test models can be represented as sketch-extrude sequences using the supported CadQuery primitives (line, arc, circle, rect, box, cylinder).
    The paper states that revolution, fillet, chamfer, and B-spline features are not supported; CC3D contains such models, so this assumption limits the scope.
  • domain assumption A pre-trained LLM's exposure to Python code transfers to CadQuery code generation after fine-tuning.
    This is the core motivation in Sections 1 and 3.1; no controlled experiment isolates the contribution of LLM pretraining.
  • domain assumption The procedurally generated training distribution is representative enough to improve performance on real test datasets.
    Ablations show empirical improvement, but this is a distribution-shift assumption, not a proven guarantee.
  • ad hoc to paper Chamfer distance reranking at inference using the input point cloud is a fair inference procedure.
    Section 4.3; this selection step may inflate CD and IR metrics relative to baselines that do not use equivalent reranking.

how reviews work

0 comments
Cite this review

Pith. "Pith review of CAD-Recode: Reverse Engineering CAD Code from Point Clouds." pith.science (2026). https://pith.science/paper/P227LUHT

@misc{pith2026241214042,
  author       = {Pith},
  title        = {Pith review of: CAD-Recode: Reverse Engineering CAD Code from Point Clouds},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P227LUHT}},
  note         = {Machine review of arXiv:2412.14042}
}
read the original abstract

Computer-Aided Design (CAD) models are typically constructed by sequentially drawing parametric sketches and applying CAD operations to obtain a 3D model. The problem of 3D CAD reverse engineering consists of reconstructing the sketch and CAD operation sequences from 3D representations such as point clouds. In this paper, we address this challenge through novel contributions across three levels: CAD sequence representation, network design, and training dataset. In particular, we represent CAD sketch-extrude sequences as Python code. The proposed CAD-Recode translates a point cloud into Python code that, when executed, reconstructs the CAD model. Taking advantage of the exposure of pre-trained Large Language Models (LLMs) to Python code, we leverage a relatively small LLM as a decoder for CAD-Recode and combine it with a lightweight point cloud projector. CAD-Recode is trained on a procedurally generated dataset of one million CAD sequences. CAD-Recode significantly outperforms existing methods across the DeepCAD, Fusion360 and real-world CC3D datasets. Furthermore, we show that our CAD Python code output is interpretable by off-the-shelf LLMs, enabling CAD editing and CAD-specific question answering from point clouds.

Figures

Figures reproduced from arXiv: 2412.14042 by the authors.

Figure 1
Figure 1. 3D CAD reverse engineering is a process of convert [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Sketch-extrude sequence (top) in DeepCAD representa [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Examples of procedurally generated CAD models. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (15 more)
Figure 4
Figure 4. Figure 4: Overview of CAD-Recode. The pipeline comprises two parts: (1) a point cloud projector (marked blue) (2) a fine-tuned pre￾trained LLM (yellow). An input point cloud is processed using (1), and outputs are then passed to an LLM (2), which predicts a CAD sketch-extrude se…
Figure 5
Figure 5. Figure 5: Qualitative results on the DeepCAD, Fusion360, and CC3D datasets. For each input point cloud (first row), we compare CAD [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Interactive editing of a CAD model. Given the code [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Examples from our procedurally generated training dataset. Each row contains CadQuery Python code and a corresponding CAD [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Our 1 M procedurally generated training dataset displays distributions CAD models that are skewed towards models with larger edge and face count per model than the DeepCAD dataset (160 k models). and significantly higher performance on the parameter accu￾racy. This dem…
Figure 9
Figure 9. Figure 9: Example models from real-world CC3D dataset. The scans exhibits numerous artifacts such as surface noise, missing parts and [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]
Figure 10
Figure 10. Figure 10: Qualitative results on the DeepCAD dataset. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png]
Figure 11
Figure 11. Figure 11: Qualitative results on the Fusion360 dataset. [PITH_FULL_IMAGE:figures/full_fig_p013_11.png]
Figure 12
Figure 12. Figure 12: Qualitative results on the real-world CC3D dataset. [PITH_FULL_IMAGE:figures/full_fig_p013_12.png]
Figure 13
Figure 13. Figure 13: CAD-Recode predictions on DeepCAD (top 2 rows), Fusion360 (mid 3 rows), and CC3D (last row) datasets. Each row contains predicted CadQuery Python code and its result after execution in Python interpreter. 14 [PITH_FULL_IMAGE:figures/full_fig_p014_13.png]
Figure 14
Figure 14. Figure 14: Examples of invalid predictions. Each row contains the ground-truth CAD model (left) and an invalid predicted CadQuery [PITH_FULL_IMAGE:figures/full_fig_p015_14.png]
Figure 15
Figure 15. Figure 15: CAD-Recode predictions from different point cloud sampling on DeepCAD, Fusion360, and real-world CC3D datasets. For each prediction, 256 points are sampled randomly from the input point cloud. Examine the following CAD code carefully to understand the 3D object it gen…
Figure 16
Figure 16. Figure 16: Example of a CAD question from SGP-Bench in the DeepCAD representation (b) and using the CadQuery representation (c). [PITH_FULL_IMAGE:figures/full_fig_p017_16.png]
Figure 17
Figure 17. Figure 17: Example of Point cloud CAD-QA (a) and answers provided by PointLLM (b), CADSIGNet and GPT-4o (c) and [PITH_FULL_IMAGE:figures/full_fig_p018_17.png]
Figure 18
Figure 18. Figure 18: Editing pipeline: given a predicted code from [PITH_FULL_IMAGE:figures/full_fig_p019_18.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

64 extracted references · 57 canonical work pages

  1. [34]

    Draw step by step: Reconstructing cad construction sequences from point clouds via multimodal diffusion

    Weijian Ma, Shuaiqi Chen, Yunzhong Lou, Xueyang Li, and Xiangdong Zhou. Draw step by step: Reconstructing cad construction sequences from point clouds via multimodal diffusion. In IEEE Conf. Comput. Vis. Pattern Recog., pages 27154–27163, 2024. 1, 2, 3, 4, 6, 12

    work page 2024

  2. [21]

    Cad-signet: Cad language inference from point clouds using layer-wise sketch instance guided attention

    Mohammad Sadil Khan, Elona Dupont, Sk Aziz Ali, Kseniya Cherenkova, Anis Kacem, and Djamila Aouada. Cad-signet: Cad language inference from point clouds using layer-wise sketch instance guided attention. In IEEE Conf. Comput. Vis. Pattern Recog., pages 4713–4722, 2024. 1, 2, 3, 4, 5, 6, 7, 8, 10, 11

    work page 2024

  3. [12]

    Transcad: A hi- erarchical transformer for cad sequence inference from point clouds

    Elona Dupont, Kseniya Cherenkova, Dimitrios Mallis, Gleb Gusev, Anis Kacem, and Djamila Aouada. Transcad: A hi- erarchical transformer for cad sequence inference from point clouds. In Eur. Conf. Comput. Vis., 2024. 2, 3, 4, 6, 11

    work page 2024

  4. [1]

    Generating CAD code with vision-language models for 3d designs

    Kamel Alrashedy, Pradyumna Tambwekar, Zulfiqar Haider Zaidi, Megan Langwasser, Wei Xu, and Matthew Gombolay. Generating CAD code with vision-language models for 3d designs. In ICLR, 2025. 3

    work page 2025

  5. [2]

    Query2cad: Generating cad models using natural language queries

    Akshay Badagabettu, Sai Sravan Yarlagadda, and Amir Barati Farimani. Query2cad: Generating cad models using natural language queries. arXiv preprint arXiv:2406.00144, 2024. 3

    arXiv 2024

  6. [3]

    Comparing 3d cad models: uses, methods, tools and per- spectives

    Antoine Bri `ere-Cˆot´e, Louis Rivest, and Roland Maranzana. Comparing 3d cad models: uses, methods, tools and per- spectives. Computer-Aided Design and Applications , 9(6): 771–794, 2012. 1

    work page 2012

  7. [4]

    Para- metric cad modeling: An analysis of strategies for design reusability

    Jorge D Camba, Manuel Contero, and Pedro Company. Para- metric cad modeling: An analysis of strategies for design reusability. Computer-Aided Design, 74:18–31, 2016. 1

    work page 2016

  8. [5]

    Mark Chen, Jerry Tworek, Heewoo Jun, Qiming Yuan, Hen- rique Ponde de Oliveira Pinto, Jared Kaplan, Harri Ed- wards, Yuri Burda, Nicholas Joseph, Greg Brockman, Alex Ray, Raul Puri, Gretchen Krueger, Michael Petrov, Heidy Khlaaf, Girish Sastry, Pamela Mishkin, Brooke Chan, Scott Gray, Nick Ryder, Mikhail Pavlov, Alethea Power, Lukasz Kaiser, Mohammad Bava...

    work page

Show all 64 references
  1. [6]

    Ll3da: Visual interactive instruction tuning for omni-3d understand- ing reasoning and planning

    Sijin Chen, Xin Chen, Chi Zhang, Mingsheng Li, Gang Yu, Hao Fei, Hongyuan Zhu, Jiayuan Fan, and Tao Chen. Ll3da: Visual interactive instruction tuning for omni-3d understand- ing reasoning and planning. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Re...

    2024

  2. [7]

    Pvdeconv: Point-voxel deconvolution for autoencoding cad construction in 3d

    Kseniya Cherenkova, Djamila Aouada, and Gleb Gusev. Pvdeconv: Point-voxel deconvolution for autoencoding cad construction in 3d. In ICIP, pages 2741–2745, 2020. 2, 4, 9

    2020

  3. [8]

    Sepicnet: Sharp edges recovery by parametric inference of curves in 3d shapes

    Kseniya Cherenkova, Elona Dupont, Anis Kacem, Ilya Arzhannikov, Gleb Gusev, and Djamila Aouada. Sepicnet: Sharp edges recovery by parametric inference of curves in 3d shapes. In CVPRW, pages 2726–2734, 2023. 3

    2023

  4. [9]

    Spelsnet: Surface primitive el- ements segmentation by b-rep graph structure supervision

    Kseniya Cherenkova, Elona Dupont, Anis Kacem, Gleb Gu- sev, and Djamila Aouada. Spelsnet: Surface primitive el- ements segmentation by b-rep graph structure supervision. NeurIPS, 2024. 3

    2024

  5. [10]

    Cadquery: A python paramet- ric cad scripting framework

    CADQuery Developers. Cadquery: A python paramet- ric cad scripting framework. https://cadquery. readthedocs.io/, 2024. Accessed: 2024-10-22. 2, 3, 4, 9

    2024

  6. [11]

    Cadops-net: Jointly learning cad operation types and steps from boundary-representations

    Elona Dupont, Kseniya Cherenkova, Anis Kacem, Sk Aziz Ali, Ilya Aryhannikov, Gleb Gusev, and Djamila Aouada. Cadops-net: Jointly learning cad operation types and steps from boundary-representations. 3DV, 2022. 3

    2022

  7. [13]

    Simultaneous registration of multiple range views for use in reverse engineering of cad models

    David W Eggert, Andrew W Fitzgibbon, and Robert B Fisher. Simultaneous registration of multiple range views for use in reverse engineering of cad models. Computer Vision and Image Understanding, 69(3):253–272, 1998. 1

    1998

  8. [14]

    Optimizing evolutionary csg tree extraction

    Markus Friedrich, Pierre-Alain Fayolle, Thomas Gabor, and Claudia Linnhoff-Popien. Optimizing evolutionary csg tree extraction. In Proceedings of the Genetic and Evolutionary Computation Conference, pages 1183–1191, 2019. 3

    2019

  9. [15]

    Complexgen: Cad reconstruction by b-rep chain complex generation

    Haoxiang Guo, Shilin Liu, Hao Pan, Yang Liu, Xin Tong, and Baining Guo. Complexgen: Cad reconstruction by b-rep chain complex generation. ACM TOG, 41(4):1–18, 2022. 3

    2022

  10. [16]

    Onellm: One framework to align all modalities with language

    Jiaming Han, Kaixiong Gong, Yiyuan Zhang, Jiaqi Wang, Kaipeng Zhang, Dahua Lin, Yu Qiao, Peng Gao, and Xi- angyu Yue. Onellm: One framework to align all modalities with language. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 26584–...

    2024

  11. [17]

    A case study on use of 3d scanning for reverse engineering and quality control

    Robin H Helle and Hirpa G Lemu. A case study on use of 3d scanning for reverse engineering and quality control. Materials Today: Proceedings, 45:5255–5262, 2021. 1

    2021

  12. [18]

    3d-llm: Inject- ing the 3d world into large language models.NeurIPS, 2023

    Yining Hong, Haoyu Zhen, Peihao Chen, Shuhong Zheng, Yilun Du, Zhenfang Chen, and Chuang Gan. 3d-llm: Inject- ing the 3d world into large language models.NeurIPS, 2023. 2

    2023

  13. [19]

    3d-llm: In- jecting the 3d world into large language models

    Yining Hong, Haoyu Zhen, Peihao Chen, Shuhong Zheng, Yilun Du, Zhenfang Chen, and Chuang Gan. 3d-llm: In- jecting the 3d world into large language models. Advances in Neural Information Processing Systems, 36:20482–20494,

  14. [20]

    Ucsg-net-unsupervised discovering of constructive solid ge- ometry tree

    Kacper Kania, Maciej Zieba, and Tomasz Kajdanowicz. Ucsg-net-unsupervised discovering of constructive solid ge- ometry tree. Adv. Neural Inform. Process. Syst., 2020. 3

    2020

  15. [22]

    Abc: A big cad model dataset for geometric deep learning

    Sebastian Koch, Albert Matveev, Zhongshi Jiang, Francis Williams, Alexey Artemov, Evgeny Burnaev, Marc Alexa, Denis Zorin, and Daniele Panozzo. Abc: A big cad model dataset for geometric deep learning. In IEEE Conf. Comput. Vis. Pattern Recog., pages 9601–9611, 2019. 2, 4

    2019

  16. [23]

    Reparamcad: Zero-shot cad re-parameterization for interactive manipula- tion

    Milin Kodnongbua, Benjamin Jones, Maaz Bin Safeer Ah- mad, Vladimir Kim, and Adriana Schulz. Reparamcad: Zero-shot cad re-parameterization for interactive manipula- tion. In SIGGRAPH Asia 2023 Conference Papers , New York, NY , USA, 2023. Association for Computing Machin- ery. 3

    2023

  17. [24]

    Brepnet: A topological message passing system for solid models

    Joseph G Lambourne, Karl DD Willis, Pradeep Kumar Jayaraman, Aditya Sanghi, Peter Meltzer, and Hooman 20 Shayani. Brepnet: A topological message passing system for solid models. In IEEE Conf. Comput. Vis. Pattern Recog., pages 12773–12782, 2021. 3, 5

    2021

  18. [25]

    Reconstructing editable prismatic cad from rounded voxel models

    Joseph George Lambourne, Karl Willis, Pradeep Kumar Ja- yaraman, Longfei Zhang, Aditya Sanghi, and Kamal Rahimi Malekshan. Reconstructing editable prismatic cad from rounded voxel models. In SIGGRAPH Asia , pages 1–9,

  19. [26]

    Supervised fitting of geometric prim- itives to 3d point clouds

    Lingxiao Li, Minhyuk Sung, Anastasia Dubrovina, Li Yi, and Leonidas J Guibas. Supervised fitting of geometric prim- itives to 3d point clouds. In IEEE Conf. Comput. Vis. Pattern Recog., pages 2652–2660, 2019. 3

    2019

  20. [27]

    Detect- ing design intent in approximate cad models using symme- try

    Ming Li, Frank C Langbein, and Ralph R Martin. Detect- ing design intent in approximate cad models using symme- try. Computer-Aided Design, 42(3):183–201, 2010. 1

    2010

  21. [28]

    Secad-net: Self-supervised cad reconstruction by learning sketch-extrude operations

    Pu Li, Jianwei Guo, Xiaopeng Zhang, and Dong-Ming Yan. Secad-net: Self-supervised cad reconstruction by learning sketch-extrude operations. In IEEE Conf. Comput. Vis. Pat- tern Recog., pages 16816–16826, 2023. 2, 3

    2023

  22. [29]

    Sfmcad: Unsupervised cad reconstruction by learning sketch-based feature modeling operations

    Pu Li, Jianwei Guo, Huibin Li, Bedrich Benes, and Dong- Ming Yan. Sfmcad: Unsupervised cad reconstruction by learning sketch-based feature modeling operations. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4671–4680, 2024. 3

    2024

  23. [30]

    Fast industrial product design method and its appli- cation based on 3d cad system

    Fei Liu. Fast industrial product design method and its appli- cation based on 3d cad system. Computer-Aided Design and Applications, 18(S3):118–128, 2020. 1

    2020

  24. [31]

    Openshape: Scaling up 3d shape representation towards open-world understanding

    Minghua Liu, Ruoxi Shi, Kaiming Kuang, Yinhao Zhu, Xu- anlin Li, Shizhong Han, Hong Cai, Fatih Porikli, and Hao Su. Openshape: Scaling up 3d shape representation towards open-world understanding. Advances in neural information processing systems, 36, 2024. 2

    2024

  25. [32]

    Pc2wf: 3d wireframe reconstruction from raw point clouds

    Yujia Liu, Stefano D’Aronco, Konrad Schindler, and Jan Dirk Wegner. Pc2wf: 3d wireframe reconstruction from raw point clouds. ICLR, 2021. 3

    2021

  26. [33]

    Multicad: Contrastive representation learning for multi-modal 3d computer-aided design models

    Weijian Ma, Minyang Xu, Xueyang Li, and Xiangdong Zhou. Multicad: Contrastive representation learning for multi-modal 3d computer-aided design models. In CIKM, New York, NY , USA, 2023. Association for Computing Ma- chinery. 2, 3, 4, 6

    2023

  27. [35]

    Sharp challenge 2023: Solving cad history and parameters recovery from point clouds and 3d scans

    Dimitrios Mallis, Ali Sk Aziz, Elona Dupont, Kseniya Cherenkova, Ahmet Serdar Karadeniz, Mohammad Sadil Khan, Anis Kacem, Gleb Gusev, and Djamila Aouada. Sharp challenge 2023: Solving cad history and parameters recovery from point clouds and 3d scans. overview, datasets, metri...

    2023

  28. [36]

    Gpt-4 technical report, 2024

    OpenAI. Gpt-4 technical report, 2024. 2, 3, 4, 6, 8, 16

    2024

  29. [37]

    Pythonocc - 3d cad/bim/plm/cam frame- work, 2022

    Thomas Paviot. Pythonocc - 3d cad/bim/plm/cam frame- work, 2022. 9

    2022

  30. [38]

    Can large language models understand symbolic graphics programs? arXiv preprint arXiv:2408.08313, 2024

    Zeju Qiu, Weiyang Liu, Haiwen Feng, Zhen Liu, Tim Z Xiao, Katherine M Collins, Joshua B Tenenbaum, Adrian Weller, Michael J Black, and Bernhard Sch¨olkopf. Can large language models understand symbolic graphics programs? arXiv preprint arXiv:2408.08313, 2024. 3, 7, 16

    2024 arXiv

  31. [39]

    Extrudenet: Unsupervised inverse sketch- and-extrude for shape parsing

    Daxuan Ren, Jianmin Zheng, Jianfei Cai, Jiatong Li, and Junzhe Zhang. Extrudenet: Unsupervised inverse sketch- and-extrude for shape parsing. In Eur. Conf. Comput. Vis., pages 482–498. Springer, 2022. 2, 3

    2022

  32. [40]

    Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter

    V Sanh. Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter. arXiv preprint arXiv:1910.01108, 2019. 3

    1910 arXiv

  33. [41]

    Vitruvion: A generative model of parametric cad sketches

    Ari Seff, Wenda Zhou, Nick Richardson, and Ryan P Adams. Vitruvion: A generative model of parametric cad sketches. In ICLR, 2022. 3

    2022

  34. [42]

    Parsenet: A parametric surface fitting network for 3d point clouds

    Gopal Sharma, Difan Liu, Subhransu Maji, Evangelos Kalogerakis, Siddhartha Chaudhuri, and Radom ´ır M ˇech. Parsenet: A parametric surface fitting network for 3d point clouds. In Eur. Conf. Comput. Vis., pages 261–276. Springer,

  35. [43]

    Industrial inspection and reverse en- gineering

    Tarek M Sobh, J Owen, C Jaynes, Mohamed Dekhil, and Thomas C Henderson. Industrial inspection and reverse en- gineering. In Proceedings of 1994 IEEE 2nd CAD-Based Vision Workshop, pages 228–235. IEEE, 1994. 1

    1994

  36. [44]

    Qwen2-1.5b

    Qwen Team. Qwen2-1.5b. https://huggingface. co/Qwen/Qwen2-1.5B, 2024. Accessed: Nov. 2024. 9

    2024

  37. [45]

    Point2cyl: Reverse engineering 3d objects from point clouds to extrusion cylinders

    Mikaela Angelina Uy, Yen-Yu Chang, Minhyuk Sung, Purvi Goel, Joseph G Lambourne, Tolga Birdal, and Leonidas J Guibas. Point2cyl: Reverse engineering 3d objects from point clouds to extrusion cylinders. In IEEE Conf. Comput. Vis. Pattern Recog., pages 11850–11860, 2022. 1, 2, 3, 6

    2022

  38. [46]

    Pie-net: Parametric inference of point cloud edges

    Xiaogang Wang, Yuelang Xu, Kai Xu, Andrea Tagliasac- chi, Bin Zhou, Ali Mahdavi-Amiri, and Hao Zhang. Pie-net: Parametric inference of point cloud edges. Adv. Neural In- form. Process. Syst., 33:20167–20178, 2020. 3

    2020

  39. [47]

    Fusion 360 gallery: A dataset and environ- ment for programmatic cad construction from human design sequences

    Karl DD Willis, Yewen Pu, Jieliang Luo, Hang Chu, Tao Du, Joseph G Lambourne, Armando Solar-Lezama, and Wo- jciech Matusik. Fusion 360 gallery: A dataset and environ- ment for programmatic cad construction from human design sequences. ACM TOG, 40(4):1–24, 2021. 2, 3, 4, 6, 11

    2021

  40. [48]

    Deepcad: A deep generative network for computer-aided design models

    Rundi Wu, Chang Xiao, and Changxi Zheng. Deepcad: A deep generative network for computer-aided design models. In IEEE Conf. Comput. Vis. Pattern Recog. , pages 6772– 6782, 2021. 2, 3, 4, 6, 7, 8, 9, 10, 11, 12, 16

    2021

  41. [49]

    Pointllm: Empowering large lan- guage models to understand point clouds

    Runsen Xu, Xiaolong Wang, Tai Wang, Yilun Chen, Jiang- miao Pang, and Dahua Lin. Pointllm: Empowering large lan- guage models to understand point clouds. InECCV, 2024. 2, 8

    2024

  42. [50]

    Inferring cad modeling sequences using zone graphs

    Xianghao Xu, Wenzhe Peng, Chin-Yi Cheng, Karl DD Willis, and Daniel Ritchie. Inferring cad modeling sequences using zone graphs. In IEEE Conf. Comput. Vis. Pattern Recog., pages 6062–6070, 2021. 2, 3

    2021

  43. [51]

    Skexgen: Autoregressive generation of cad con- struction sequences with disentangled codebooks

    Xiang Xu, Karl DD Willis, Joseph G Lambourne, Chin- Yi Cheng, Pradeep Kumar Jayaraman, and Yasutaka Fu- rukawa. Skexgen: Autoregressive generation of cad con- struction sequences with disentangled codebooks. In ICML, pages 24698–24724. PMLR, 2022. 3, 9

    2022

  44. [52]

    Hierarchical neural coding for controllable cad model generation

    Xiang Xu, Pradeep Kumar Jayaraman, Joseph G Lambourne, Karl DD Willis, and Yasutaka Furukawa. Hierarchical neural coding for controllable cad model generation. ICML, 2023. 3, 6, 12 21

    2023

  45. [53]

    Ulip: Learning a unified representation of language, images, and point clouds for 3d understanding

    Le Xue, Mingfei Gao, Chen Xing, Roberto Mart ´ın-Mart´ın, Jiajun Wu, Caiming Xiong, Ran Xu, Juan Carlos Niebles, and Silvio Savarese. Ulip: Learning a unified representation of language, images, and point clouds for 3d understanding. In Proceedings of the IEEE/CVF conference o...

    2023

  46. [54]

    Ulip-2: Towards scalable multimodal pre-training for 3d understand- ing

    Le Xue, Ning Yu, Shu Zhang, Artemis Panagopoulou, Jun- nan Li, Roberto Mart´ın-Mart´ın, Jiajun Wu, Caiming Xiong, Ran Xu, Juan Carlos Niebles, and Silvio Savarese. Ulip-2: Towards scalable multimodal pre-training for 3d understand- ing. In Proceedings of the IEEE/CVF Conferenc...

    2024

  47. [55]

    Qwen2 technical report, 2024

    An Yang, Baosong Yang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Zhou, Chengpeng Li, Chengyuan Li, Dayiheng Liu, Fei Huang, Guanting Dong, Haoran Wei, Huan Lin, Jialong Tang, Jialin Wang, Jian Yang, Jianhong Tu, Jian- wei Zhang, Jianxin Ma, Jianxin Yang, Jin Xu, Jingren Zhou, Jin...

    2024

  48. [56]

    Shapegpt: 3d shape generation with a unified multi-modal language model, 2023

    Fukun Yin, Xin Chen, Chi Zhang, Biao Jiang, Zibo Zhao, Ji- ayuan Fan, Gang Yu, Taihao Li, and Tao Chen. Shapegpt: 3d shape generation with a unified multi-modal language model, 2023. 2

    2023

  49. [57]

    Img2cad: Reverse engineering 3d cad models from images through vlm-assisted conditional factorization

    Yang You, Mikaela Angelina Uy, Jiaqi Han, Rahul Thomas, Haotong Zhang, Suya You, and Leonidas Guibas. Img2cad: Reverse engineering 3d cad models from images through vlm-assisted conditional factorization. arXiv preprint arXiv:2408.01437, 2024. 3

    2024

  50. [58]

    D2csg: Unsupervised learning of compact csg trees with dual complements and dropouts

    Fenggen Yu, Qimin Chen, Maham Tanveer, Ali Mah- davi Amiri, and Hao Zhang. D2csg: Unsupervised learning of compact csg trees with dual complements and dropouts. Advances in Neural Information Processing Systems , 36: 22807–22819, 2023. 3

    2023

  51. [59]

    D2csg: Unsupervised learning of compact csg trees with dual complements and dropouts

    Fenggen Yu, Qimin Chen, Maham Tanveer, Ali Mah- davi Amiri, and Hao Zhang. D2csg: Unsupervised learning of compact csg trees with dual complements and dropouts. Advances in Neural Information Processing Systems , 36,

  52. [60]

    Cadtalk: An algorithm and benchmark for semantic commenting of cad programs

    Haocheng Yuan, Jing Xu, Hao Pan, Adrien Bousseau, Niloy J Mitra, and Changjian Li. Cadtalk: An algorithm and benchmark for semantic commenting of cad programs. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3753–3762, 2024. 3

    2024

  53. [61]

    Pointclip: Point cloud understanding by clip

    Renrui Zhang, Ziyu Guo, Wei Zhang, Kunchang Li, Xu- peng Miao, Bin Cui, Yu Qiao, Peng Gao, and Hongsheng Li. Pointclip: Point cloud understanding by clip. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8552–8562, 2022. 2

    2022

  54. [62]

    Michelangelo: Conditional 3d shape generation based on shape-image-text aligned latent representation

    Zibo Zhao, Wen Liu, Xin Chen, Xianfang Zeng, Rui Wang, Pei Cheng, BIN FU, Tao Chen, Gang YU, and Shenghua Gao. Michelangelo: Conditional 3d shape generation based on shape-image-text aligned latent representation. In Thirty- seventh Conference on Neural Information Processing ...

    2023

  55. [63]

    Nerve: Neural volumetric edges for parametric curve extraction from point cloud

    Xiangyu Zhu, Dong Du, Weikai Chen, Zhiyou Zhao, Yinyu Nie, and Xiaoguang Han. Nerve: Neural volumetric edges for parametric curve extraction from point cloud. In IEEE Conf. Comput. Vis. Pattern Recog. , pages 13601–13610,

  56. [64]

    Point- clip v2: Prompting clip and gpt for powerful 3d open-world learning

    Xiangyang Zhu, Renrui Zhang, Bowei He, Ziyu Guo, Ziyao Zeng, Zipeng Qin, Shanghang Zhang, and Peng Gao. Point- clip v2: Prompting clip and gpt for powerful 3d open-world learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 2639–2650,

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.