Research map / Image generation / Text-to-image diffusion
Arbitrary-Shaped Image Generation via Spherical Neural Field Diffusion
Jiyuan Xia, Yuanshen Guan, Ruikang Xu, Zhiwei Xiong
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 |
|---|---|---|---|
| add spherical residual blocks (ablation) | ASIG · Matterport3D panorama generation | FID: 27.93 -> 25.49 | Improved |
| add spherical residual blocks (ablation) | ASIG · Matterport3D panorama reconstruction | PSNR: 28.65 -> 30.07 | Improved |
| add spherical residual blocks (ablation) | ASIG · Matterport3D panorama reconstruction | LPIPS: 0.2205 -> 0.1680 | Improved |
| add Seam-Aware Padding in the spherical neural field (ablation) | ASIG · Matterport3D panorama generation | FID: 27.69 -> 25.49 | Improved |
| add Seam-Aware Padding in the spherical neural field (ablation) | ASIG · Matterport3D panorama reconstruction | PSNR: 29.17 -> 30.07 | Improved |
| add Seam-Aware Padding in the diffusion UNet (ablation) | ASIG · panorama generation | visual seams | Improved |
| use spherical neural field instead of LTEW enhancement (ablation) | ASIG vs SMGD+ · panorama at 1536 resolution | pFID: 27.31 -> 16.31 | Improved |
| use spherical neural field instead of LTEW enhancement (ablation) | ASIG vs SDXL+ · panorama at 512 resolution | pFID: 59.84 -> 20.49 | Improved |
| use mesh-based spherical representation instead of ERP | ASIG · panorama reconstruction | PSNR: 29.07 -> 30.07 | Improved |
| use mesh-based spherical representation instead of cubemap | ASIG · panorama reconstruction | PSNR: 29.42 -> 30.07 | Improved |
| add LSDIR perspective images to neural field training | ASIG · panorama generation | FID: 25.49 -> 25.22 | Improved |
| add Seam-Aware Padding to UNet | ASIG · UNet inference time | time (s): 3.23 -> 3.62 | Hurt |
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.
- MVDiffusion — MVDiffusion: Enabling Holistic Multi-view Image Generation with Correspondence-Aware Diffusion (NeurIPS 2023) · better on 3 of 3 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
- Neural Point Cloud Diffusion for Disentangled 3D Shape and Appearance Generation (CVPR 2024 · similarity 0.87)
- SphereDiffusion: Spherical Geometry-Aware Distortion Resilient Diffusion Model (AAAI 2024 · similarity 0.86)
- Denoising Diffusion via Image-Based Rendering (ICLR 2024 · similarity 0.86)
- SIGNeRF: Scene Integrated Generation for Neural Radiance Fields (CVPR 2024 · similarity 0.86)
- A 3D Scene is Worth 1K Tokens: 3D-Grounded Representation for Scene Generation at Scale (NeurIPS 2026 · similarity 0.85)
- Efficient-3Dim: Learning a Generalizable Single-image Novel-view Synthesizer in One Day (ICLR 2024 · similarity 0.85)
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.
