Research map / Image generation / Autoregressive and latent generation
Thinking Image Color Aesthetics Assessment: Models, Datasets and Benchmarks
Shuai He, Anlong Ming, Yaqi Li, Jinyuan Sun, ShunTian Zheng, Huadong 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 Delegate Transformer instead of the best prior IAA model | ICAA17K · ICAA17K | SRCC: .855 -> .873 | Improved |
| add multi-task training with holistic aesthetics score (ablation) | Delegate Transformer · ICAA17K | SRCC: .873 -> .887 | Improved |
| place Delegate blocks in last two stages instead of Swin (ablation) | Delegate Transformer · ICAA17K | SRCC: 0.830 -> 0.873 | Improved |
| place Delegate blocks in last two stages instead of DAT deformable attention (ablation) | Delegate Transformer · ICAA17K | SRCC: 0.836 -> 0.873 | Improved |
| use Delegate blocks in early stages (ablation) | Delegate Transformer · ICAA17K | SRCC: 0.873 -> 0.828 | Hurt |
| use Delegate blocks in all four stages (ablation) | Delegate Transformer · ICAA17K | SRCC: 0.873 -> 0.837 | Hurt |
| too few or too many interest points (vary Neti kernel size) (ablation) | Delegate Transformer · ICAA17K | SRCC | Hurt |
| set number of color spaces kappa to 7 (ablation) | Delegate Transformer · ICAA17K | SRCC | Improved |
| train on ICAA17K instead of SPAQ | baseline IAA models · ICAA17K vs SPAQ | SRCC and LCC | Improved |
Closest papers by meaning
- Thinking Aesthetics Assessment of Image Color Temperature: Models, Datasets and Benchmarks (AAAI 2026 · similarity 0.90)
- Towards Artistic Image Aesthetics Assessment: A Large-Scale Dataset and a New Method (CVPR 2023 · similarity 0.88)
- ArtiMuse: Fine-Grained Image Aesthetics Assessment with Joint Scoring and Expert-Level Understanding (CVPR 2026 · similarity 0.87)
- Advancing Comprehensive Aesthetic Insight with Multi-Scale Text-Guided Self-Supervised Learning (AAAI 2025 · similarity 0.85)
- ELTA: An Enhancer against Long-Tail for Aesthetics-oriented Models (ICML 2024 · similarity 0.84)
- Bridging Cognitive Gap: Hierarchical Description Learning for Artistic Image Aesthetics Assessment (AAAI 2026 · 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.
