Research map / 3D vision / Face and shape
TAPS3D: Text-Guided 3D Textured Shape Generation From Pseudo Supervision
Jiacheng Wei, Hao Wang, Jiashi Feng, Guosheng Lin, Kim-Hui Yap
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 |
|---|---|---|---|
| Use TAPS3D instead of CLIP-NeRF | Text-guided 3D generation · Car | FID: 67.8 -> 20.1 | Improved |
| Use TAPS3D instead of CLIP-NeRF | Text-guided 3D generation · Chair | FID: 48.4 -> 43.7 | Improved |
| Use TAPS3D instead of TITG3SG | Text-guided 3D generation · Chair | CLIP R-Precision: 9.05 ± 1.92 -> 20.19 ± 1.89 | Improved |
| Use TAPS3D instead of TITG3SG | Text-guided 3D generation · Table | CLIP R-Precision: 6.67 ± 3.25 -> 15.21 ± 3.15 | Improved |
| Use TAPS3D instead of CLIP-Forge | 3D geometry generation · Chair | FPD: 825.96 -> 342.23 | Improved |
| Use TAPS3D instead of TITG3SG | 3D geometry generation · Table | FPD: 1639.68 -> 1468.43 | Improved |
| Use pseudo-text CLIP features rather than image CLIP features as training input (ablation) | TAPS3D · ShapeNet Car, Chair, Table, Motorbike | FID: Car: 68.7 -> 21.7; Chair: 81.1 -> 44.8; Table: 94.3 -> 43.2; Motorbike: 146.5 -> 57.2 | Improved |
| Add CLIP loss (ablation) | TAPS3D · ShapeNet Car | FID: 83.7 -> 21.7 | Improved |
| Add CLIP loss (ablation) | TAPS3D · ShapeNet Chair | FID: 148.9 -> 44.8 | Improved |
| Add image regularization loss (ablation) | TAPS3D · ShapeNet Car | FID: 59.8 -> 21.7 | Improved |
| Add image regularization loss (ablation) | TAPS3D · ShapeNet Chair | FID: 144.2 -> 44.8 | Improved |
| Add background augmentation (ablation) | TAPS3D · ShapeNet Car | FID: 55.1 -> 21.7 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CLIP-NeRF — CLIP-NeRF: Text-and-Image Driven Manipulation of Neural Radiance Fields (CVPR 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Towards Implicit Text-Guided 3D Shape Generation (CVPR 2022 · similarity 0.88)
- PI3D: Efficient Text-to-3D Generation with Pseudo-Image Diffusion (CVPR 2024 · similarity 0.88)
- Dream3D: Zero-Shot Text-to-3D Synthesis Using 3D Shape Prior and Text-to-Image Diffusion Models (CVPR 2023 · similarity 0.88)
- ShapeWords: Guiding Text-to-Image Synthesis with 3D Shape-Aware Prompts (CVPR 2025 · similarity 0.87)
- ViewDiff: 3D-Consistent Image Generation with Text-to-Image Models (CVPR 2024 · similarity 0.87)
- TextField3D: Towards Enhancing Open-Vocabulary 3D Generation with Noisy Text Fields (ICLR 2024 · similarity 0.87)
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.
