Research map / Image generation / Autoregressive and latent generation
MetaIQA: Deep Meta-Learning for No-Reference Image Quality Assessment
Check 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 |
|---|---|---|---|
| Learn a quality-prior initialization by bi-level meta-training across distortion-specific tasks, then fine-tune for the target task | ResNet18 NR-IQA · TID2013 leave-one-distortion-out (average) | SROCC: 0.7073 -> 0.8539 | Helped |
| Learn a quality-prior initialization by bi-level meta-training across distortion-specific tasks, then fine-tune for the target task | ResNet18 NR-IQA · KADID-10K leave-one-distortion-out (average) | SROCC: 0.6295 -> 0.7672 | Helped |
| Transfer the synthetic-distortion meta-model and fine-tune on authentic-distortion training data | MetaIQA · CID2013 | PLCC: 0.825 NSSADNN; 0.784 MetaIQA | Mixed |
| Transfer the synthetic-distortion meta-model and fine-tune on authentic-distortion training data | MetaIQA · CID2013 | SROCC: 0.748 -> 0.766 | Helped |
| Transfer the synthetic-distortion meta-model and fine-tune on authentic-distortion training data | MetaIQA · LIVE challenge | PLCC: 0.813 -> 0.835 | Helped |
| Transfer the synthetic-distortion meta-model and fine-tune on authentic-distortion training data | MetaIQA · LIVE challenge | SROCC: 0.745 -> 0.802 | Helped |
| Transfer the synthetic-distortion meta-model and fine-tune on authentic-distortion training data | MetaIQA · KonIQ-10K | PLCC: 0.813 -> 0.887 | Helped |
| Transfer the synthetic-distortion meta-model and fine-tune on authentic-distortion training data | MetaIQA · KonIQ-10K | SROCC: 0.805 -> 0.850 | Helped |
| Add bi-level meta-training instead of direct Adam training on distortion-specific images (ablation) | ResNet18 NR-IQA · CID2013 | PLCC: 0.727 -> 0.784 | Helped |
| Add bi-level meta-training instead of direct Adam training on distortion-specific images (ablation) | ResNet18 NR-IQA · CID2013 | SROCC: 0.712 -> 0.766 | Helped |
| Add bi-level meta-training instead of direct Adam training on distortion-specific images (ablation) | ResNet18 NR-IQA · LIVE challenge | PLCC: 0.801 -> 0.835 | Helped |
| Add bi-level meta-training instead of direct Adam training on distortion-specific images (ablation) | ResNet18 NR-IQA · LIVE challenge | SROCC: 0.743 -> 0.802 | Helped |
Closest papers by meaning
- Transformer-Based No-Reference Image Quality Assessment via Supervised Contrastive Learning (AAAI 2024 · similarity 0.91)
- Adaptive Feature Selection for No-Reference Image Quality Assessment by Mitigating Semantic Noise Sensitivity (ICML 2024 · similarity 0.88)
- Boosting Image Quality Assessment through Efficient Transformer Adaptation with Local Feature Enhancement (CVPR 2024 · similarity 0.88)
- Causal-IQA: Towards the Generalization of Image Quality Assessment Based on Causal Inference (ICML 2024 · similarity 0.88)
- Toward Generalized Image Quality Assessment: Relaxing the Perfect Reference Quality Assumption (CVPR 2025 · similarity 0.87)
- Adaptive Image Quality Assessment via Teaching Large Multimodal Model to Compare (NeurIPS 2024 · similarity 0.87)
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.