Research map / Image generation / Autoregressive and latent generation
Content-Variant Reference Image Quality Assessment via Knowledge Distillation
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 |
|---|---|---|---|
| add non-aligned reference images without distillation (ablation) | NR-student baseline · TID2013 and KonIQ-10K | SRCC: 0.631 -> 0.679; 0.317 -> 0.352 | Helped |
| add knowledge distillation from the FR-teacher (ablation) | NAR-student · TID2013 and KonIQ-10K | SRCC: 0.679 -> 0.691; 0.352 -> 0.416 | Helped |
| add knowledge distillation from the FR-teacher (ablation) | NAR-student · KonIQ-10K, across different NAR references | SRCC standard deviation: 0.072 -> 0.004 | Helped |
| use pixel-aligned instead of content-variant reference images (ablation) | NAR-student · LIVE, CSIQ, and TID2013 | SRCC: 0.913 -> 0.958; 0.829 -> 0.937; 0.691 -> 0.846 | Helped |
| use content-similar instead of content-variant NAR references (ablation) | NAR-student · LIVE, CSIQ, and TID2013 | SRCC: 0.913 -> 0.931; 0.829 -> 0.862; 0.691 -> 0.720 | Helped |
| use higher-quality HQ reference images (ablation) | NAR-student · TID2013 and KonIQ-10K | SRCC: 0.665 to 0.691; 0.392 to 0.416 | Helped |
| increase the number and size of input patches (ablation) | CVRKD-IQA FR-teacher and NAR-student · TID2013 | SRCC | Helped |
| use content-variant non-aligned references with FR-teacher distillation | CVRKD-IQA · LIVE, CSIQ, TID2013, and KonIQ-10K | SRCC: 0.913; 0.829; 0.691; 0.416 | Helped |
Closest papers by meaning
- Toward Generalized Image Quality Assessment: Relaxing the Perfect Reference Quality Assumption (CVPR 2025 · similarity 0.92)
- Adaptive Image Quality Assessment via Teaching Large Multimodal Model to Compare (NeurIPS 2024 · similarity 0.90)
- Distilling Spatially-Heterogeneous Distortion Perception for Blind Image Quality Assessment (CVPR 2025 · similarity 0.88)
- Revisiting MLLM Based Image Quality Assessment: Errors and Remedy (AAAI 2026 · similarity 0.88)
- Adaptive Feature Selection for No-Reference Image Quality Assessment by Mitigating Semantic Noise Sensitivity (ICML 2024 · similarity 0.87)
- Augmenting Perceptual Super-Resolution via Image Quality Predictors (CVPR 2025 · similarity 0.86)
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.