NOFACE3D: NOise-scheduled Focused Anchoring and Controlled Enhancement for 3D Generation

Key points

• Recent image-to-3D models generate complete, high-quality assets from a single image, with global shape and appearance that are already convincing; what remains is the fine scale, where small, thin, elongated, or detail-rich regions are reconstructed softly and under-resolved.1
• Given a generated mesh, its input image, and a refinement box, NOFACE3D extracts the selected submesh, scales it up to fill the generator's input volume, and re-encodes it into the full-resolution latent grid.1
• We implement NOFACE3D on TRELLIS.2 and its sparse voxel representation; all of it operates purely at inference time on the frozen pretrained generator.1
• We evaluate NOFACE3D on PAniC-3D, TexVerse and BigDetail, a new benchmark of 339 assets with annotated detail-critical regions.1

Sources (1)

• [1]NOFACE3D: NOise-scheduled Focused Anchoring and Controlled Enhancement for 3D Generation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 9, 06:05 AM

  To this end, we introduce NOFACE3D, a training-free framework that refines these detail-critical regions while preserving the rest of the asset.

  Recent image-to-3D models generate complete, high-quality assets from a single image, with global shape and appearance that are already convincing; what remains is the fine scale, where small, thin, elongated, or detail-rich regions are reconstructed softly and under-resolved.

Extractive summary: sentences quoted from the sources.