Research map / Image generation / Autoregressive and latent generation
Very Long Natural Scenery Image Prediction by Outpainting
Zongxin Yang, Jian Dong, Ping Liu, Yi Yang, Shuicheng Yan
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 Global Residual Blocks (3 instead of 0) (ablation) | RCT+SHC outpainting network · Scenery dataset | FID: 15.171 -> 13.713 | Improved |
| add Global Residual Blocks (3 instead of 0) (ablation) | RCT+SHC outpainting network · Scenery dataset | IS: 2.756 -> 2.852 | Improved |
| use 1 GRB instead of 0 (ablation) | RCT+SHC outpainting network · Scenery dataset | FID: 15.171 -> 14.828 | Improved |
| use RCT instead of fully connected layer (ablation) | SHC outpainting network · Scenery dataset | FID: 15.186 -> 13.713 | Improved |
| use RCT instead of fully connected layer (ablation) | SHC outpainting network · Scenery dataset | IS: 2.845 -> 2.852 | Improved |
| add Skip Horizontal Connection layers (ablation) | encoder-decoder outpainting network · Scenery dataset | boundary smoothness (qualitative) | Improved |
| add Recurrent Content Transfer to the network (ablation) | encoder-decoder outpainting network · Scenery dataset multi-step prediction | consistency and sharpness (qualitative) | Improved |
| use contextual attention (CA) instead of RCT+SHC | outpainting baseline · Scenery dataset | IS vs FID: IS 2.931 vs 2.852; FID 19.040 vs 13.713 | Mixed |
Closest papers by meaning
- Continuous-Multiple Image Outpainting in One-Step via Positional Query and A Diffusion-based Approach (ICLR 2024 · similarity 0.89)
- EasyOutPainter: One Step Image Outpainting With Both Continuous Multiple and Resolution (TPAMI 2025 · similarity 0.89)
- Infinite-Canvas: Higher-Resolution Video Outpainting with Extensive Content Generation (AAAI 2025 · similarity 0.85)
- Unboxed: Geometrically and Temporally Consistent Video Outpainting (CVPR 2025 · similarity 0.85)
- Autoregressive Omni-Aware Outpainting for Open-Vocabulary 360-Degree Image Generation (AAAI 2024 · similarity 0.84)
- Shadow-Enlightened Image Outpainting (CVPR 2024 · similarity 0.84)
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.
