Research map / Image generation / Autoregressive and latent generation
Re-IQA: Unsupervised Learning for Image Quality Assessment in the Wild
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 |
|---|---|---|---|
| Combine content-aware and quality-aware representations instead of using either alone (ablation) | Re-IQA · KonIQ | SRCC: 0.896 -> 0.914; 0.861 -> 0.914 | Helped |
| Combine content-aware and quality-aware representations instead of either alone (ablation) | Re-IQA · FLIVE | SRCC: 0.588 -> 0.645; 0.584 -> 0.645 | Helped |
| Combine content-aware and quality-aware representations instead of either alone (ablation) | Re-IQA · LIVE-IQA | SRCC: 0.867 -> 0.970; 0.971 -> 0.970 | Mixed |
| Use n_aug=11 rather than 2, 5, 15, or 23 (ablation) | Re-IQA with content and quality modules · KonIQ | SRCC: 0.895, 0.901, 0.914, 0.898, 0.895 | Helped |
| Use patch size 160 rather than 128, 192, 224, or 256 (ablation) | Re-IQA with content and quality modules · KonIQ | SRCC: 0.910, 0.914, 0.911, 0.905, 0.897 | Helped |
| Use OLA bound 10-30% rather than 5-15%, 50-80%, or no bound (ablation) | Re-IQA with content and quality modules · KonIQ | SRCC: 0.903, 0.914, 0.897, 0.905 | Helped |
| Use n_aug=11 rather than 2, 5, 15, or 23 (ablation) | Re-IQA with content and quality modules · SPAQ | SRCC: 0.902, 0.913, 0.918, 0.910, 0.909 | Mixed |
| Use patch size 160 rather than 128, 192, 224, or 256 (ablation) | Re-IQA with content and quality modules · CSIQ-IQA | SRCC: 0.940, 0.947, 0.942, 0.937, 0.932 | Helped |
| Use OLA bound 10-30% rather than 5-15%, 50-80%, or no bound (ablation) | Re-IQA quality-aware module · SPAQ | SRCC: 0.889, 0.900, 0.886, 0.890 | Helped |
Closest papers by meaning
- Adaptive Image Quality Assessment via Teaching Large Multimodal Model to Compare (NeurIPS 2024 · similarity 0.89)
- Boosting Image Quality Assessment through Efficient Transformer Adaptation with Local Feature Enhancement (CVPR 2024 · similarity 0.89)
- Adaptive Feature Selection for No-Reference Image Quality Assessment by Mitigating Semantic Noise Sensitivity (ICML 2024 · similarity 0.88)
- Toward Generalized Image Quality Assessment: Relaxing the Perfect Reference Quality Assumption (CVPR 2025 · similarity 0.88)
- Augmenting Perceptual Super-Resolution via Image Quality Predictors (CVPR 2025 · similarity 0.88)
- Teaching Large Language Models to Regress Accurate Image Quality Scores Using Score Distribution (CVPR 2025 · similarity 0.88)
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.