mlsanityData to 7 Oct 2026

MeshFIM: Local Low-Poly Mesh Editing via Fill-in-the-Middle Autoregressive Generation

arXiv 2605.08744 ↗First posted 9 May 2026cs.GR cs.AI cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Autoregressive (AR) models can generate high-quality low-poly meshes from point clouds, but they still operate in an all-or-nothing manner: when a local region is unsatisfactory, the entire mesh must be regenerated, wasting computation and destroying satisfactory mesh structure elsewhere. We introduce MeshFIM, a Fill-in-the-Middle (FIM) framework that regenerates a target region of a low-poly mesh conditioned on the surrounding context. MeshFIM addresses three mesh-specific challenges: enforcing exact attachment along the exposed boundary, preserving topological order in the context, and suppressing overflow beyond the intended region. It does so with five complementary design choices: boundary vertex markers, context positional embeddings, expanded context width, context augmentation, and a low-poly geometry encoder whose gated subtraction mechanism focuses generation on the missing region by leveraging the difference between the reference surface and the existing mesh. Detailed ablation studies are presented to show the effectiveness of every introduced component. Based on MeshFIM, we demonstrate two applications: interactive brush-based editing and automatic defect repair on low-poly mesh (see Figure 1). Last but not least, experiments show that MeshFIM outperforms a range of baselines in mesh refinement, mesh repair and whole mesh generation plus stitch-back scheme.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 27 of its 40 references that are in the map. First posted on arXiv 9 May 2026.

Builds on

27 of 40 references.

References · 27

  1. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,874
  2. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,880
  3. 2023Code Llama: Open Foundation Models for Codecited by 2,105
  4. 2022Objaverse: A Universe of Annotated 3D Objectscited by 1,522
  5. 2024DeepSeek-Coder: When the Large Language Model Meets Programming -- The Rise of Code Intelligencecited by 1,154
  6. 2023Objaverse-XL: A Universe of 10M+ 3D Objectscited by 767
  7. 2021Perceiver IO: A General Architecture for Structured Inputs & Outputscited by 616
  8. 2022InCoder: A Generative Model for Code Infilling and Synthesiscited by 492
  9. 20233DShape2VecSet: A 3D Shape Representation for Neural Fields and Generative Diffusion Modelscited by 449
  10. 2020PolyGen: An Autoregressive Generative Model of 3D Meshescited by 267
  11. 2023MeshGPT: Generating Triangle Meshes with Decoder-Only Transformerscited by 254
  12. 2022Efficient Training of Language Models to Fill in the Middlecited by 248
15 more
  1. 2023Michelangelo: Conditional 3D Shape Generation based on Shape-Image-Text Aligned Latent Representationcited by 225
  2. 2024MeshAnything: Artist-Created Mesh Generation with Autoregressive Transformerscited by 140
  3. 2024MeshAnything V2: Artist-Created Mesh Generation With Adjacent Mesh Tokenizationcited by 102
  4. 2024MeshXL: Neural Coordinate Field for Generative 3D Foundation Modelscited by 100
  5. 2021Hierarchical Transformers Are More Efficient Language Modelscited by 99
  6. 2024Meshtron: High-Fidelity, Artist-Like 3D Mesh Generation at Scalecited by 87
  7. 2024EdgeRunner: Auto-regressive Auto-encoder for Artistic Mesh Generationcited by 84
  8. 2024Scaling Mesh Generation via Compressive Tokenizationcited by 71
  9. 2025DeepMesh: Auto-Regressive Artist-mesh Creation with Reinforcement Learningcited by 62
  10. 2020Neural Subdivisioncited by 49
  11. 2025TreeMeshGPT: Artistic Mesh Generation with Autoregressive Tree Sequencingcited by 38
  12. 2019Fast Tetrahedral Meshing in the Wildcited by 35
  13. 2025ARMesh: Autoregressive Mesh Generation via Next-Level-of-Detail Predictioncited by 11
  14. 2025QuadGPT: Native Quadrilateral Mesh Generation with Autoregressive Modelscited by 11
  15. 2023Learning Self-Prior for Mesh Inpainting Using Self-Supervised Graph Convolutional Networkscited by 2

Built on by

Not cited yet.