InteriorFusion / README.md
stevee00's picture
Upload README.md
0a1186d verified

InteriorFusion: Single Image β†’ High-Quality Editable 3D Interior Scenes

InteriorFusion is a production-grade, open-source AI system for converting a single 2D interior photograph into a high-quality, editable 3D scene with textured meshes, Gaussian splats, and PBR materials.

πŸ† Why InteriorFusion?

Current image-to-3D pipelines (TRELLIS, Hunyuan3D-2, TripoSR, Stable Fast 3D) are designed for single objects, not rooms. InteriorFusion is the first system specifically architected for interior design and room reconstruction.

Key Differentiators

  • Scene-aware: Understands room layouts, wall/floor topology, and spatial relationships
  • Multi-object coherent: Furniture doesn't float; objects respect real-world scale
  • Editable: Individual furniture objects are separable, movable, replaceable
  • PBR materials: Metallic, roughness, normal maps for photorealistic rendering
  • Multi-format export: GLB, FBX, OBJ, USDZ, Gaussian Splatting
  • Under 30s generation on RTX 4090

Architecture Overview

Single Interior Image
       β”‚
       β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Phase 1: Scene Understanding        β”‚
β”‚ - Depth Anything V2 (metric indoor) β”‚
β”‚ - Room layout estimation             β”‚
β”‚ - Semantic segmentation             β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
       β”‚
       β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Phase 2: Multi-View Generation       β”‚
β”‚ - Zero123++ / SyncDreamer            β”‚
β”‚ - Depth-conditioned view synthesis   β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
       β”‚
       β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Phase 3: 3D Reconstruction           β”‚
β”‚ - Room shell (walls/floor/ceiling)    β”‚
β”‚ - Per-object Gaussian reconstruction β”‚
β”‚ - TRELLIS.2 for furniture detail    β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
       β”‚
       β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Phase 4: Scene Assembly             β”‚
β”‚ - SpatialLM layout optimization      β”‚
β”‚ - Collision detection & correction   β”‚
β”‚ - Scale normalization               β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
       β”‚
       β–Ό
β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”
β”‚ Phase 5: Material & Texture         β”‚
β”‚ - PBR material generation            β”‚
β”‚ - Texture baking                     β”‚
β”‚ - Light estimation                   β”‚
β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
       β”‚
       β–Ό
  Export: GLB / FBX / OBJ / USDZ / 3DGS

Model Zoo

Model HF Path Purpose Size
InteriorFusion-S stevee00/InteriorFusion-S Fast preview, 512Β³ 1.5B
InteriorFusion-L stevee00/InteriorFusion-L Production quality, 1024Β³ 4B
InteriorFusion-XL stevee00/InteriorFusion-XL Research quality, 2048Β³ 10B

Quick Start

pip install -e .
python -m interiorfusion.infer --image room_photo.jpg --output ./output/

License: MIT