Research map / 3D vision / 3D generation
StyleFusion360: View-Consistent Head Stylization via Adaptive Style Modulation
Furkan Guzelant, Arda Goktogan, Tarık Kaya, Aysegul Dundar
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| modulate values instead of keys in Style Fusion Attention (ablation) | StyleFusion360 · Style modulation ablation | FID: 71.685 -> 84.602 | Hurt |
| modulate both keys and values instead of keys only (ablation) | StyleFusion360 · Style modulation ablation | FID: 71.685 -> 88.220 | Hurt |
| modulate values instead of keys only (ablation) | StyleFusion360 · Style modulation ablation | depth L2 difference: 0.0234 -> 0.0253 | Hurt |
| modulate values instead of keys only (ablation) | StyleFusion360 · Style modulation ablation | ID similarity: 0.4245 -> 0.4881 | Improved |
| modulate both keys and values instead of keys only (ablation) | StyleFusion360 · Style modulation ablation | CLIP similarity: 0.7939 -> 0.7968 | Improved |
| add AdaIN to key modulation (ablation) | StyleFusion360 · FFHQ/RenderMe360 (qualitative) | style transfer | Improved |
| add attention to key modulation (ablation) | StyleFusion360 · FFHQ/RenderMe360 (qualitative) | spatial consistency | Improved |
| integrate StyleAligned into the 3D-aware pipeline (ablation) | StyleFusion360 · FFHQ (qualitative) | multi-view consistency | Hurt |
| apply same modulation to value features in addition to keys (ablation) | StyleFusion360 · FFHQ (qualitative) | visual quality | No clear effect |
| use StyleFusion360 instead of Identity3DHead | Identity3DHead · FFHQ | FID: 88.26 -> 64.48 | Improved |
| use StyleFusion360 instead of DiffusionGAN3D | DiffusionGAN3D · FFHQ | MEt3R: 9.68 -> 9.99 | Mixed |
| use StyleFusion360 instead of StyleID | StyleID · RenderMe360 | FID: 90.58 -> 74.86 | Improved |
Closest papers by meaning
- Styl3R: Instant 3D Stylized Reconstruction for Arbitrary Scenes and Styles (NeurIPS 2025 · similarity 0.89)
- Stylos: Multi-View 3D Stylization with Single-Forward Gaussian Splatting (ICLR 2026 · similarity 0.89)
- 3DToonify: Creating Your High-Fidelity 3D Stylized Avatar Easily from 2D Portrait Images (CVPR 2024 · similarity 0.87)
- StyleMaster: Stylize Your Video with Artistic Generation and Translation (CVPR 2025 · similarity 0.86)
- Morpheus: Text-Driven 3D Gaussian Splat Shape and Color Stylization (CVPR 2025 · similarity 0.86)
- DiffPortrait360: Consistent Portrait Diffusion for 360 View Synthesis (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
