Research map / 3D vision / Depth and stereo
Learning a Deep Color Difference Metric for Photographic Images
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 a multi-scale autoregressive normalizing flow feature transform with Euclidean distance | CD-Flow · SPCD all pairs | STRESS: 21.431 -> 18.473 | Helped |
| use a multi-scale autoregressive normalizing flow feature transform with Euclidean distance | CD-Flow · SPCD all pairs | PLCC: 0.846 -> 0.871 | Helped |
| use a multi-scale autoregressive normalizing flow feature transform with Euclidean distance | CD-Flow · SPCD all pairs | SRCC: 0.842 -> 0.865 | Helped |
| use the learned CD-Flow metric rather than CD-Net under mild geometric distortions | CD-Flow · SPCD translation, rotation, and dilation | SRCC: 0.842 -> 0.856; 0.772 -> 0.816; 0.773 -> 0.797 | Helped |
| use 2 flow steps instead of 8 (ablation) | CD-Flow · SPCD all pairs | STRESS: 18.473 -> 21.633 | Hurt |
| use 4 flow steps instead of 8 (ablation) | CD-Flow · SPCD all pairs | STRESS: 18.473 -> 20.994 | Hurt |
| use 16 flow steps instead of 8 (ablation) | CD-Flow · SPCD all pairs | STRESS: 18.473 -> 17.792 | Helped |
| use 2 scales instead of 6 (ablation) | CD-Flow · SPCD all pairs | STRESS: 18.473 -> 24.686 | Hurt |
| use 4 scales instead of 6 (ablation) | CD-Flow · SPCD all pairs | STRESS: 18.473 -> 19.730 | Hurt |
| use 8 scales instead of 6 (ablation) | CD-Flow · SPCD all pairs | STRESS: 18.473 -> 18.524 | No clear effect |
Closest papers by meaning
- DreamSim: Learning New Dimensions of Human Visual Similarity using Synthetic Data (NeurIPS 2023 · similarity 0.86)
- MegaCoin: Enhancing Medium-Grained Color Perception for Vision-Language Models (AAAI 2026 · similarity 0.84)
- Logarithmic Lenses: Exploring Log RGB Data for Image Classification (CVPR 2024 · similarity 0.84)
- Close Imitation of Expert Retouching for Black-and-White Photography (CVPR 2024 · similarity 0.84)
- Color Alignment in Diffusion (CVPR 2025 · similarity 0.83)
- ColorConceptBench: A Benchmark for Probabilistic Color-Concept Understanding in Text-to-Image Models (NeurIPS 2026 · 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.