Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
MiCADangelo: Fine-Grained Reconstruction of Constrained CAD Models from 3D Scans
Authors: Ahmet Karadeniz, Dimitrios Mallis, Danila Rukhovich, Kseniya Cherenkova, Anis Kacem, Djamila Aouada
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we detail the experimental setup and report results to evaluate the effectiveness of the proposed Mi CADangelo. Datasets. We evaluate our method on the test sets of Deep CAD [6] and Fusion360 [18] datasets. ... Metrics. For 3D CAD reconstruction, median Chamfer Distance, Intersection over Union (Io U), median Edge Chamfer Distance (ECD) and Invalidity Ratio (IR) are used. For 2D sketches, image-level sketch chamfer distance (SCD) is used. Additional metric details are provided in the supplementary. |
| Researcher Affiliation | Collaboration | Ahmet Serdar Karadeniz Sn T, University of Luxembourg EMAIL Dimitrios Mallis Sn T, University of Luxembourg EMAIL Danila Rukhovich Sn T, University of Luxembourg EMAIL Kseniya Cherenkova Sn T, University of Luxembourg, Artec 3D EMAIL Anis Kacem Sn T, University of Luxembourg EMAIL Djamila Aouada Sn T, University of Luxembourg EMAIL |
| Pseudocode | No | The paper describes the proposed method in Section 4 and its subsections, along with illustrative diagrams (Figure 1 and Figure 2), but it does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks or figures. |
| Open Source Code | No | The data used in this study is publicly available. Scripts for preprocessing the data and extracting additional information used in our work will also be made publicly accessible. However, the code for the method and training cannot currently be released under an open-source license that permits use, modification, and distribution, due to constraints related to the industrial collaboration under which this work was conducted. Nonetheless, all necessary details for reproducing the results are thoroughly provided in the paper and the accompanying supplementary materials. |
| Open Datasets | Yes | Datasets. We evaluate our method on the test sets of Deep CAD [6] and Fusion360 [18] datasets. The sketch plane detection is trained on the train set of Deep CAD [6], while the constrained sketch parameterization network is trained on the train set of the Sketch Graphs [35] dataset and finetuned on an augmented version of the dataset. ... Justification: The data used in this study is publicly available. |
| Dataset Splits | Yes | Datasets. We evaluate our method on the test sets of Deep CAD [6] and Fusion360 [18] datasets. The sketch plane detection is trained on the train set of Deep CAD [6], while the constrained sketch parameterization network is trained on the train set of the Sketch Graphs [35] dataset and finetuned on an augmented version of the dataset. |
| Hardware Specification | No | The paper mentions that "The paper clearly states the compute resources used for the experiments and the compute time of the method in Section 5 and the supplementary material" in its NeurIPS checklist justification. However, the main text of the paper provided does not specify concrete hardware details such as GPU/CPU models, memory amounts, or processor types used for running its experiments. |
| Software Dependencies | No | Implementation Details. ... Adam W is used as an optimizer for all the experiments. More details are in the supplementary. While an optimizer (Adam W) is mentioned, no specific version numbers for software components (e.g., Python, PyTorch, CUDA) are provided. |
| Experiment Setup | Yes | Implementation Details. Planes are preprocessed to ensure consistent normal directions along the positive axes. The plane detection model is trained for 20 epochs on Deep CAD [6] with lr = 1 10 4. The constrained sketch parameterization model is trained on Sketch Graphs [35] as in [37] and finetuned for 50 epochs on synthetically generated, noise-augmented loops. We use 40 cross-sections per axis, each normalized to an image of size 128 128. Normalization is performed using 2D offsets and a scale factor corresponding to a unit bounding box. The transformer encoder of the sketch plane detection comprises 4 layers and 4 attention heads, with an embedding dimension of 256. The sketch parameterization network architecture is similar to [37]. For the extrusion optimization, we run 200 iterations with a learning rate of 2 10 4. Adam W is used as an optimizer for all the experiments. More details are in the supplementary. |