Research map / Image generation / Image restoration
Measuring Perceptual Color Differences of Smartphone Photographs
Zhihua Wang, Keshuo Xu, Yang Yang, Jianlei Dong, Shuhang Gu, Lihao Xu, Yuming Fang, 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 |
|---|---|---|---|
| add 11x11 conv to a 1x1-only front-end filter bank (ablation) | CD-Net · SPCD | STRESS: 24.756 -> 21.431 | Improved |
| add 1x1 conv to an 11x11-only front-end filter bank (ablation) | CD-Net · SPCD | STRESS: 21.932 -> 21.431 | Improved |
| increase last conv channels from 3 to 18 (ablation) | CD-Net · SPCD | STRESS: 22.289 -> 21.261 | No clear effect |
| increase last conv channels from 3 to 12 (ablation) | CD-Net · SPCD | PLCC: 0.843 -> 0.846 | No clear effect |
| train on larger images, 256 vs 768 (ablation) | CD-Net · SPCD (test 256) | STRESS: 21.789 -> 21.643 | Improved |
| increase test image size from 256 to 1024 with training at 256 (ablation) | CD-Net · SPCD | STRESS: 21.789 -> 22.291 | Hurt |
| learn CD-Net on SPCD vs spatial filtering extension S-CIELAB of CIEDE2000 | CIEDE2000 · SPCD | STRESS: 31.439 -> 32.780 | Hurt |
| replace CIEDE2000 with learned CD-Net | CIEDE2000 · SPCD | PLCC: 0.726 -> 0.846 | Improved |
| replace CIELAB with CD-Net on unseen color patches | CIELAB · COM dataset | STRESS: 45.202 -> 38.872 | Improved |
| replace CIEDE2000 with CD-Net on color patches | CIEDE2000 · COM dataset | STRESS: 28.979 -> 38.872 | Hurt |
| replace CD-Net with FLIP on color distortions | CD-Net · TID2013 subset | STRESS: 15.474 -> 12.610 | Hurt |
| replace CD-Net with deep CNN backbone ResNet-18 (more parameters) | CD-Net · SPCD | STRESS: 21.431 -> 18.574 | Mixed |
Closest papers by meaning
- Multiscale Sliced Wasserstein Distances as Perceptual Color Difference Measures (ECCV 2024 · similarity 0.91)
- Learning a Deep Color Difference Metric for Photographic Images (CVPR 2023 · similarity 0.90)
- Perceptual Quality Assessment of Smartphone Photography (CVPR 2020 · similarity 0.88)
- Thinking Aesthetics Assessment of Image Color Temperature: Models, Datasets and Benchmarks (AAAI 2026 · similarity 0.86)
- Thinking Image Color Aesthetics Assessment: Models, Datasets and Benchmarks (ICCV 2023 · similarity 0.85)
- DiffRAW: Leveraging Diffusion Model to Generate DSLR-Comparable Perceptual Quality sRGB from Smartphone RAW Images (AAAI 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.
