Research map / Architectures and efficiency / Attention and transformers
Outpainting by Queries
Kai Yao, Penglei Gao, Xi Yang, Jie Sun, Rui Zhang, Kaizhu Huang
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 pretrained ViT encoder (ablation) | QueryOTR · Scenery | FID: 22.784 -> 20.366 | Improved |
| increase decoder layers from 4 to 8 (ablation) | QueryOTR · Scenery | FID: 20.366 -> 20.373 | No clear effect |
| reduce decoder layers from 4 to 2 (ablation) | QueryOTR · Scenery | FID: 20.366 -> 20.731 | Hurt |
| add reconstruction loss (ablation) | QueryOTR · Scenery | FID: 31.282 -> 20.366 | Improved |
| add perceptual loss (ablation) | QueryOTR · Scenery | FID: 33.380 -> 20.366 | Improved |
| add Query Expansion Module over learnable queries (ablation) | QueryOTR · Scenery | FID: 36.967 -> 22.784 | Improved |
| add noise to QEM padded tokens (ablation) | QueryOTR · Scenery | FID: 23.444 -> 22.784 | Improved |
| add deformable convolution in QEM (ablation) | QueryOTR · Scenery | FID: 23.530 -> 22.784 | Improved |
| add per-patch normalization (no PSM) (ablation) | QueryOTR · Scenery | FID: 51.945 -> 31.073 | Improved |
| add PSM (with per-patch norm) (ablation) | QueryOTR · Scenery | FID: 31.073 -> 20.366 | Improved |
| add PSM (without per-patch norm) (ablation) | QueryOTR · Scenery | FID: 51.945 -> 22.501 | Improved |
| add QEM to pure transformer for convergence (ablation) | QueryOTR · Scenery | training convergence speed: 3.3 times faster | Improved |
Closest papers by meaning
- Continuous-Multiple Image Outpainting in One-Step via Positional Query and A Diffusion-based Approach (ICLR 2024 · similarity 0.88)
- Reconstructing the Image Stitching Pipeline: Integrating Fusion and Rectangling into a Unified Inpainting Model (NeurIPS 2024 · similarity 0.84)
- SyFormer: Structure-Guided Synergism Transformer for Large-Portion Image Inpainting (AAAI 2024 · similarity 0.84)
- LookHere: Vision Transformers with Directed Attention Generalize and Extrapolate (NeurIPS 2024 · similarity 0.84)
- Shadow-Enlightened Image Outpainting (CVPR 2024 · similarity 0.84)
- Unboxed: Geometrically and Temporally Consistent Video Outpainting (CVPR 2025 · 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.
