Research map / Image generation / Fast diffusion sampling
Target-Balanced Score Distillation
Zhou Xu, Qi Wang, Yuxiao Yang, Luyuan Zhang, Zhang Liang, Yang Li
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 |
|---|---|---|---|
| add target negative prompts (TNP) in place of general negative prompts to Bridge (ablation) | Bridge · Text-to-3D qualitative | texture quality vs shape | Mixed |
| add target-balanced dynamic weighting (TBSD) on top of TNP (ablation) | TNP-based score distillation · Text-to-3D qualitative | shape and texture quality | Improved |
| use TBSD instead of SDI (multi-step) for 3D generation | text-to-3D · Text-to-3D (43 prompts) | CLIP score: 33.47 -> 33.59 | Improved |
| use TBSD instead of SDS | text-to-3D · Text-to-3D (43 prompts) | CLIP score: 29.81 -> 33.59 | Improved |
| vary the specific wording of negative prompts in TNP (six GPT-4 variants) (ablation) | TBSD · Text-to-3D qualitative | visual quality | No clear effect |
| increase alpha and beta (stronger, longer shape bias) (ablation) | TBSD · Text-to-3D qualitative | texture detail and color | Hurt |
| increase fixed coefficient a of the classifier-free guidance term in TNP (ablation) | TNP optimization · Text-to-3D qualitative | shape vs texture | Mixed |
Closest papers by meaning
- Rethinking Score Distillation as a Bridge Between Image Distributions (NeurIPS 2024 · similarity 0.90)
- Rethinking Score Distilling Sampling for 3D Editing and Generation (ICML 2025 · similarity 0.90)
- Taming Mode Collapse in Score Distillation for Text-to-3D Generation (CVPR 2024 · similarity 0.89)
- Text-to-3D with Classifier Score Distillation (ICLR 2024 · similarity 0.89)
- Score Distillation via Reparametrized DDIM (NeurIPS 2024 · similarity 0.89)
- Consistent3D: Towards Consistent High-Fidelity Text-to-3D Generation with Deterministic Sampling Prior (CVPR 2024 · similarity 0.89)
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.
