Research map / Image generation / Autoregressive and latent generation
Multiscale Sliced Wasserstein Distances as Perceptual Color Difference Measures
Jiaqi He, Zhihua Wang, Leon Wang, Tsein-I Liu, Yuming Fang, Qilin Sun, Kede Ma
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 non-local patch distribution matching (MS-SWD) instead of co-located CIEDE2000 | color difference measure · SPCD non-perfectly aligned pairs | PLCC: 0.653 -> 0.841 | Improved |
| use MS-SWD instead of CIEDE2000 on perfectly aligned pairs | color difference measure · SPCD perfectly aligned pairs | PLCC: 0.827 -> 0.778 | Hurt |
| horizontally flip one image of each pair (ablation) | MS-SWD · SPCD flipping | PLCC: 0.841 -> 0.836 | No clear effect |
| horizontally flip one image of each pair (ablation) | CIELAB · SPCD flipping | PLCC: 0.683 -> 0.168 | Hurt |
| increase random projections from 4 to 128 (ablation) | MS-SWD · SPCD non-perfectly aligned pairs | PLCC: 0.804 -> 0.841 | Improved |
| increase random projections from 128 to 256 (ablation) | MS-SWD · SPCD non-perfectly aligned pairs | PLCC: 0.841 -> 0.842 | No clear effect |
| replace random linear projections with learnable non-linear projections (ablation) | MS-SWD · SPCD all pairs | PLCC: 0.794 -> 0.884 | Improved |
| use multiple scales K instead of fewer scales (ablation) | MS-SWD · reference image recovery | pixel-level fidelity | Improved |
| use CIELAB color space instead of sRGB (ablation) | MS-SWD · SPCD | performance | Improved |
| add CSF-based spatial pre-filtering (ablation) | MS-SWD · SPCD | performance | No clear effect |
Closest papers by meaning
- Learning a Deep Color Difference Metric for Photographic Images (CVPR 2023 · similarity 0.91)
- Measuring Perceptual Color Differences of Smartphone Photographs (TPAMI 2023 · similarity 0.91)
- CD-iNet: Deep Invertible Network for Perceptual Image Color Difference Measurement (IJCV 2024 · similarity 0.85)
- Thinking Image Color Aesthetics Assessment: Models, Datasets and Benchmarks (ICCV 2023 · similarity 0.84)
- Color Conditional Generation with Sliced Wasserstein Guidance (NeurIPS 2025 · similarity 0.83)
- Color Alignment in Diffusion (CVPR 2025 · similarity 0.83)
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.
