Research map / Image generation / Autoregressive and latent generation
DR.Experts: Differential Refinement of Distortion-Aware Experts for Blind Image Quality Assessment
Bohan Fu, Guanyi Qin, Fazhan Zhang, Zihao Huang, Mingxuan Li, Runze Hu
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 |
|---|---|---|---|
| add DSDM to image encoder (ablation) | Image Encoder · KonIQ | SRCC: 0.916 -> 0.930 | Improved |
| add DSDM to image encoder (ablation) | Image Encoder · LIVEC | SRCC: 0.857 -> 0.885 | Improved |
| add DDWM on top of DSDM (ablation) | DSDM · KonIQ | SRCC: 0.930 -> 0.941 | Improved |
| add DDWM on top of DSDM (ablation) | DSDM · LIVEC | SRCC: 0.885 -> 0.914 | Improved |
| use raw DA-CLIP features alone instead of image encoder (ablation) | Image Encoder · KonIQ | SRCC: 0.916 -> 0.720 | Hurt |
| add bridging features to Dis and Sem groups (ablation) | DDWM feature group · LIVEC | SRCC: 0.909 -> 0.914 | Improved |
| add distortion features to Sem and Bri groups (ablation) | DDWM feature group · LIVEC | SRCC: 0.913 -> 0.914 | Improved |
| add semantic features to Dis and Bri groups (ablation) | DDWM feature group · LIVEC | PLCC: 0.919 -> 0.926 | Improved |
| combine all three feature groups instead of only distortion features (ablation) | DDWM feature group · KonIQ | SRCC: 0.933 -> 0.941 | Improved |
| use DR.Experts instead of LODA when trained on KonIQ and tested on LIVEC | DR.Experts · Cross-dataset | SRCC: 0.811 -> 0.841 | Improved |
| use DR.Experts instead of LODA when trained on LIVEFB and tested on LIVEC | DR.Experts · Cross-dataset | SRCC: 0.805 -> 0.807 | No clear effect |
| train with only 60%/40%/20% of data (DR.Experts vs LoDa at 20%) | DR.Experts · LIVEC (20% data) | SRCC: 0.815 -> 0.837 | Improved |
Closest papers by meaning
- Blind Image Quality Assessment via Vision-Language Correspondence: A Multitask Learning Perspective (CVPR 2023 · similarity 0.91)
- Distilling Spatially-Heterogeneous Distortion Perception for Blind Image Quality Assessment (CVPR 2025 · similarity 0.90)
- Integrating Global Context Contrast and Local Sensitivity for Blind Image Quality Assessment (ICML 2024 · similarity 0.89)
- Bridging the Synthetic-to-Authentic Gap: Distortion-Guided Unsupervised Domain Adaptation for Blind Image Quality Assessment (CVPR 2024 · similarity 0.88)
- Traversing Distortion-Perception Tradeoff using a Single Score-Based Generative Model (CVPR 2025 · similarity 0.86)
- Evaluation of Visual Processing Should Be Human-Centered, Not Metric-Centered (NeurIPS 2026 · similarity 0.85)
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.
