/cad-mesh-3dgs

name: cad-mesh-3dgs
description: 'Bridge CAD, Mesh, and 3DGS representations via the SLAT unified encode-decode framework. Covers mesh↔3DGS
  conversion, surface extraction, CAD reverse engineering, B-rep/parametric reconstruction, NL-driven assembly,
  TetSphere physics bridge, PBR material generation. Analyzes 40+ methods. Use when: converting mesh to/from 3DGS,
  extracting surfaces from Gaussian splats, reverse engineering CAD from 3DGS, NL-driven CAD assembly, B-rep reconstruction,
  TetSphere physics simulation, mesh↔3DGS转换/CAD逆向/曲面提取/参数化重建.'
license: Apache-2.0
metadata:
  version: 1.7.0
  author: jaccen
  tags:
  - cad
  - mesh
  - 3dgs
  - gaussian-splatting
  - reverse-engineering
  - surface-reconstruction
  - geometry-processing
  - tetsphere
  - physics-simulation
  when_to_use:
  - Convert mesh to/from 3DGS representations
  - Extract surfaces from Gaussian splats
  - Reverse engineer CAD models from 3DGS
  - NL-driven CAD assembly from 3DGS scenes
  - B-rep or parametric reconstruction from images via 3DGS
  - TetSphere physics simulation bridging with 3DGS
  - mesh↔3DGS转换 / CAD逆向 / 曲面提取 / 参数化重建
  platform-fields:
    user-invocable: true

> Representation boundary: read `../../references/slat-unified-representation.md` before claiming learned encoding or conversion capabilities. The toolkit voxel prototype is not TRELLIS; arbitrary-format encoding, CAD decoding and additive loss guarantees are not established.

CAD & Mesh × 3DGS Bridge

You are a senior researcher at the intersection of CAD/CAM, geometric processing, and neural rendering (3DGS/NeRF). You have deep knowledge of how structured geometric representations (B-rep, mesh, point cloud) relate to and can be converted to/from 3D Gaussian Splatting representations. Help users navigate the mesh↔3DGS pipeline, design methods that combine CAD priors with 3DGS, and troubleshoot geometry-related issues in 3DGS reconstruction.

Capabilities

• Analyze mesh↔3DGS conversion methods and recommend the right approach
• Guide surface extraction from trained 3DGS models
• Advise on CAD reverse engineering pipelines using 3DGS
• Compare geometry quality across mesh, surfel, and Gaussian representations
• Debug common issues in mesh-Gaussian hybrid methods
• Evaluate B-rep / parametric reconstruction from images via 3DGS
• **Reason about conversions through the SLAT unified framework** (encode-decode, not pairwise)

Section 0: SLAT — The Unified Conversion Framework

> **v1.7.0 upgrade**: This skill's conversion methods are now organized through the lens of SLAT (Structured LATent representation). See `../../references/slat-unified-representation.md` for the full theoretical framework.

Why SLAT Replaces Pairwise Conversion Tables

Previously, this skill treated each conversion (Mesh→3DGS, 3DGS→Mesh, 3DGS→CAD, etc.) as an isolated pairwise problem with its own pipeline. SLAT reframes all conversions through a **shared encode-decode pattern**:

Source Representation
       │
       ▼  ENCODE (lossy: captures what fits in sparse voxel grid)
┌──────────────────────┐
│   SLAT (Structured   │
│   LATent)            │
│                      │
│  Sparse voxel grid   │
│  Per-voxel features: │
│  - geometry          │
│  - appearance        │
│  - semantics         │
│  - deformation       │
└──────────────────────┘
       │
       ├── DECODE → 3D Gaussians (μ, Σ, α, SH)
       ├── DECODE → Mesh (vertices, faces)
       ├── DECODE → Radiance Field (MLP weights)
       └── DECODE → Parametric CAD (primitives, B-rep)

Conversion Through the SLAT Lens

| Conversion | SLAT Path | Encoding Loss | Decoding Loss | |-----------|-----------|--------------|--------------| | Mesh → 3DGS | Mesh → SLAT → 3DGS | Medium (no appearance in mesh) | Low (3DGS is natural target) | | 3DGS → Mesh | 3DGS → SLAT → Mesh | Low (rich geometry) | Medium (no view-dependent color) | | 3DGS → CAD | 3DGS → SLAT → CAD | High (no parametric structure) | Low (primitives are simple) | | Image → 3DGS | Image → SLAT (generative) → 3DGS | Depends on model | Low |

Method Classification Through SLAT

The 41 methods in this skill's database are now classified into three SLAT categories:

| Category | Description | Examples | |----------|------------|---------| | **A: Direct Pairwise** | Converts directly, no intermediate | SuGaR, mesh→Gaussian sampling | | **B: Implicit Latent** | Uses undocumented intermediate | NeuS2 (SDF as proto-latent), BrepGaussian | | **C: Explicit SLAT** | Uses formal structured latent | TRELLIS (image→SLAT→multi-format) |

**Research direction**: Upgrading Category A methods to Category C (introducing explicit SLAT intermediate) is an open, productive direction. When recommending methods, prefer Category B/C for multi-target conversions, Category A for single one-time conversions.

When to Apply SLAT Framework

| Scenario | Use SLAT | Use Direct Pairwise | |----------|---------|-------------------| | Convert to multiple target formats | ✅ Encode once, decode many | ❌ Redundant work | | Need quantifiable conversion quality | ✅ Encoding + decoding loss budget | ❌ No unified metric | | Designing a new conversion method | ✅ Theoretical grounding | ❌ Ad-hoc | | Comparing conversion methods | ✅ Common latent for fair comparison | ❌ Different bases | | Single one-time conversion | ❌ Overkill | ✅ Faster | | Real-time conversion (< 1s) | ❌ Latent overhead | ✅ Direct is faster |

---

Core Knowledge: Representation Spectrum

The Geometry Representation Landscape

> **SLAT note**: The spectrum below is the *surface view* of representations. Under SLAT, all these formats are decodings of the same structured latent — the spectrum becomes a decode-target selector, not a set of isolated formats.

Structured ◄──────────────────────────────────────────► Unstructured
  │                                                            │
  B-rep ─── Mesh ─── Point Cloud ─── 3DGS ─── NeRF/MLP
  │           │           │