Awesome AI4CAD Hub Papers · Datasets · Projects
← Back to papers

HiFi-BRep: High-Fidelity Latent Representation for Robust B-Rep Generation

arXiv 2026 63.1 method

TLDR

HiFi-BRep introduces a topology-aware encoder and single-stage decoder for robust high-fidelity B-Rep generation, improving structural validity and geometric fidelity over state-of-the-art methods.

Reasoning

The paper clearly identifies key limitations in existing B-Rep generation and proposes novel solutions with topology-guided attention and differentiable manifold constraints. However, the abstract lacks specific dataset details, quantitative results, and explicit real-world application context, making it hard to fully assess empirical scope.

Read-first score

Read-first score 63.1, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 75.

Recency 8%
100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Reproducibility 25%
81

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=True; dataset=False; markers=code,github

Methodology quality 25%
60

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=experiment,metric

Topical relevance 42%
46.9

Uses existing LLM keyword relevance scores normalized to 0-100. AI for CAD,computer-aided design,neural CAD,generative CAD,parametric CAD,B-Rep,boundary representation,constructive solid geometry,CSG,sketch extrusion,CAD generation,CAD reconstruction,text-to-CAD,image-to-CAD,point cloud to CAD,CAD program

Field roles

FrontierReproducibility anchor

Rank sensitivity

Stability: volatile; rank range: 15.

Keyword Scores

B-Rep
10
boundary representation
10
computer-aided design
9
generative CAD
9
CAD generation
9
neural CAD
8
AI for CAD
7
CAD reconstruction
4
parametric CAD
2
constructive solid geometry
1
CSG
1
sketch extrusion
1
text-to-CAD
1
image-to-CAD
1
point cloud to CAD
1
CAD program
1

Deep Analysis

Innovations

  • Topology-aware encoder with learnable queries to eliminate padding and topology-guided attention to prevent feature contamination
  • Single-stage decoder that jointly predicts geometry and topology in parallel, embedding manifold constraints as a differentiable learning objective

Methodology

HiFi-BRep uses a topology-aware encoder to construct a high-fidelity latent representation by eliminating padding via learnable queries and preventing feature contamination with topology-guided attention. A single-stage decoder then jointly predicts geometry and topology in parallel, embedding core manifold constraints as a differentiable learning objective to ensure mutual guidance and avoid cascaded errors.

Key Results

HiFi-BRep significantly outperforms state-of-the-art methods in both structural validity and geometric fidelity.

Tags