BrepGen: A B-rep Generative Diffusion Model with Structured Latent Geometry
2401.15563
0
2
π
Abstract
This paper presents BrepGen, a diffusion-based generative approach that directly outputs a Boundary representation (B-rep) Computer-Aided Design (CAD) model. BrepGen represents a B-rep model as a novel structured latent geometry in a hierarchical tree. With the root node representing a whole CAD solid, each element of a B-rep model (i.e., a face, an edge, or a vertex) progressively turns into a child-node from top to bottom. B-rep geometry information goes into the nodes as the global bounding box of each primitive along with a latent code describing the local geometric shape. The B-rep topology information is implicitly represented by node duplication. When two faces share an edge, the edge curve will appear twice in the tree, and a T-junction vertex with three incident edges appears six times in the tree with identical node features. Starting from the root and progressing to the leaf, BrepGen employs Transformer-based diffusion models to sequentially denoise node features while duplicated nodes are detected and merged, recovering the B-Rep topology information. Extensive experiments show that BrepGen advances the task of CAD B-rep generation, surpassing existing methods on various benchmarks. Results on our newly collected furniture dataset further showcase its exceptional capability in generating complicated geometry. While previous methods were limited to generating simple prismatic shapes, BrepGen incorporates free-form and doubly-curved surfaces for the first time. Additional applications of BrepGen include CAD autocomplete and design interpolation. The code, pretrained models, and dataset are available at https://github.com/samxuxiang/BrepGen.
Create account to get full access
Overview
- BrepGen is a diffusion-based generative approach that directly outputs a Boundary representation (B-rep) Computer-Aided Design (CAD) model.
- It represents a B-rep model as a novel structured latent geometry in a hierarchical tree, with the root node representing the whole CAD solid and each element of the B-rep model (face, edge, vertex) becoming a child-node.
- BrepGen employs Transformer-based diffusion models to sequentially denoise node features while detecting and merging duplicated nodes to recover the B-rep topology information.
- Extensive experiments show that BrepGen advances the task of CAD B-rep generation, surpassing existing methods and showcasing its ability to generate complicated geometry with free-form and doubly-curved surfaces.
Plain English Explanation
BrepGen is a new way to generate 3D CAD models using a diffusion-based approach. It represents the CAD model as a hierarchical tree, where the root node represents the entire model, and each part of the model (like a face, edge, or vertex) is represented by a child node.
The key idea is that BrepGen uses a series of Transformer-based diffusion models to gradually refine the geometry and topology of the model. It starts with a simple shape at the root and progressively adds more detail as it moves down the tree.
This allows BrepGen to generate complex CAD models, including those with curved surfaces and intricate shapes, which previous methods struggled with. The researchers show that BrepGen outperforms other 3D model generation techniques on various benchmarks, and it can be used for applications like CAD autocomplete and design interpolation.
Technical Explanation
BrepGen represents a B-rep CAD model as a hierarchical tree, with the root node representing the whole solid and each element (face, edge, vertex) becoming a child-node. The geometry information is stored in the nodes as the global bounding box and a latent code describing the local shape, while the topology is implicitly represented by node duplication.
The key innovation is how BrepGen uses Transformer-based diffusion models to generate this tree-structured representation. Starting from the root, it sequentially denoises the node features while detecting and merging duplicated nodes to recover the B-rep topology. This allows BrepGen to generate complex CAD models with free-form and doubly-curved surfaces, going beyond the limitations of previous methods that were restricted to simpler, prismatic shapes.
The researchers demonstrate BrepGen's capabilities through extensive experiments on various benchmarks, showing that it outperforms existing CAD model generation approaches. They also present results on a new furniture dataset, further showcasing BrepGen's exceptional ability to generate complicated geometries.
Critical Analysis
The paper presents a novel and promising approach to CAD model generation, but there are a few areas that could be explored further:
-
Generalization and Scalability: While BrepGen can generate complex geometries, it's unclear how well it would scale to very large or intricate CAD models. The researchers should investigate the model's performance and limitations as the complexity of the target models increases.
-
User-Interaction and Control: The current version of BrepGen is a fully-automated generation system. Incorporating user-interaction or control mechanisms, such as allowing users to guide the generation process or specify design constraints, could make the system more practical for real-world CAD design workflows.
-
Robustness and Reliability: The paper does not address potential issues around the consistency or reliability of the generated models. Assessing the model's sensitivity to input variations and ensuring the generated models are watertight and suitable for downstream CAD applications would be valuable.
-
Computational Efficiency: The computational requirements of the diffusion-based approach used in BrepGen are not discussed. Exploring ways to improve the efficiency of the generation process would make the system more practical for real-time or interactive applications.
Overall, BrepGen represents an exciting step forward in the field of 3D shape generation and CAD model synthesis. With further research and development, it could become a valuable tool for designers and engineers working with complex 3D geometries.
Conclusion
BrepGen is a novel diffusion-based approach that can directly generate high-quality Boundary Representation (B-rep) CAD models, overcoming the limitations of previous methods. By representing the B-rep model as a hierarchical tree and using Transformer-based diffusion models, BrepGen is able to generate complex geometries with free-form and doubly-curved surfaces, as demonstrated on various benchmarks and a new furniture dataset.
While the paper presents a significant advancement in the field of 3D shape generation and CAD model synthesis, there are opportunities for further research to improve the system's generalization, user-interaction, robustness, and computational efficiency. Overall, BrepGen shows great promise in transforming how designers and engineers create and interact with complex 3D models.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
Hyperbolic Geometric Latent Diffusion Model for Graph Generation
Xingcheng Fu, Yisen Gao, Yuecen Wei, Qingyun Sun, Hao Peng, Jianxin Li, Xianxian Li
0
0
Diffusion models have made significant contributions to computer vision, sparking a growing interest in the community recently regarding the application of them to graph generation. Existing discrete graph diffusion models exhibit heightened computational complexity and diminished training efficiency. A preferable and natural way is to directly diffuse the graph within the latent space. However, due to the non-Euclidean structure of graphs is not isotropic in the latent space, the existing latent diffusion models effectively make it difficult to capture and preserve the topological information of graphs. To address the above challenges, we propose a novel geometrically latent diffusion framework HypDiff. Specifically, we first establish a geometrically latent space with interpretability measures based on hyperbolic geometry, to define anisotropic latent diffusion processes for graphs. Then, we propose a geometrically latent diffusion process that is constrained by both radial and angular geometric properties, thereby ensuring the preservation of the original topological properties in the generative graphs. Extensive experimental results demonstrate the superior effectiveness of HypDiff for graph generation with various topologies.
5/7/2024
BlockFusion: Expandable 3D Scene Generation using Latent Tri-plane Extrapolation
Zhennan Wu, Yang Li, Han Yan, Taizhang Shang, Weixuan Sun, Senbo Wang, Ruikai Cui, Weizhe Liu, Hiroyuki Sato, Hongdong Li, Pan Ji
0
0
We present BlockFusion, a diffusion-based model that generates 3D scenes as unit blocks and seamlessly incorporates new blocks to extend the scene. BlockFusion is trained using datasets of 3D blocks that are randomly cropped from complete 3D scene meshes. Through per-block fitting, all training blocks are converted into the hybrid neural fields: with a tri-plane containing the geometry features, followed by a Multi-layer Perceptron (MLP) for decoding the signed distance values. A variational auto-encoder is employed to compress the tri-planes into the latent tri-plane space, on which the denoising diffusion process is performed. Diffusion applied to the latent representations allows for high-quality and diverse 3D scene generation. To expand a scene during generation, one needs only to append empty blocks to overlap with the current scene and extrapolate existing latent tri-planes to populate new blocks. The extrapolation is done by conditioning the generation process with the feature samples from the overlapping tri-planes during the denoising iterations. Latent tri-plane extrapolation produces semantically and geometrically meaningful transitions that harmoniously blend with the existing scene. A 2D layout conditioning mechanism is used to control the placement and arrangement of scene elements. Experimental results indicate that BlockFusion is capable of generating diverse, geometrically consistent and unbounded large 3D scenes with unprecedented high-quality shapes in both indoor and outdoor scenarios.
5/27/2024
Split-and-Fit: Learning B-Reps via Structure-Aware Voronoi Partitioning
Yilin Liu, Jiale Chen, Shanshan Pan, Daniel Cohen-Or, Hao Zhang, Hui Huang
0
0
We introduce a novel method for acquiring boundary representations (B-Reps) of 3D CAD models which involves a two-step process: it first applies a spatial partitioning, referred to as the ``split``, followed by a ``fit`` operation to derive a single primitive within each partition. Specifically, our partitioning aims to produce the classical Voronoi diagram of the set of ground-truth (GT) B-Rep primitives. In contrast to prior B-Rep constructions which were bottom-up, either via direct primitive fitting or point clustering, our Split-and-Fit approach is top-down and structure-aware, since a Voronoi partition explicitly reveals both the number of and the connections between the primitives. We design a neural network to predict the Voronoi diagram from an input point cloud or distance field via a binary classification. We show that our network, coined NVD-Net for neural Voronoi diagrams, can effectively learn Voronoi partitions for CAD models from training data and exhibits superior generalization capabilities. Extensive experiments and evaluation demonstrate that the resulting B-Reps, consisting of parametric surfaces, curves, and vertices, are more plausible than those obtained by existing alternatives, with significant improvements in reconstruction quality. Code will be released on https://github.com/yilinliu77/NVDNet.
6/11/2024
Bayesian Diffusion Models for 3D Shape Reconstruction
Haiyang Xu, Yu Lei, Zeyuan Chen, Xiang Zhang, Yue Zhao, Yilin Wang, Zhuowen Tu
0
0
We present Bayesian Diffusion Models (BDM), a prediction algorithm that performs effective Bayesian inference by tightly coupling the top-down (prior) information with the bottom-up (data-driven) procedure via joint diffusion processes. We show the effectiveness of BDM on the 3D shape reconstruction task. Compared to prototypical deep learning data-driven approaches trained on paired (supervised) data-labels (e.g. image-point clouds) datasets, our BDM brings in rich prior information from standalone labels (e.g. point clouds) to improve the bottom-up 3D reconstruction. As opposed to the standard Bayesian frameworks where explicit prior and likelihood are required for the inference, BDM performs seamless information fusion via coupled diffusion processes with learned gradient computation networks. The specialty of our BDM lies in its capability to engage the active and effective information exchange and fusion of the top-down and bottom-up processes where each itself is a diffusion process. We demonstrate state-of-the-art results on both synthetic and real-world benchmarks for 3D shape reconstruction.
4/23/2024