Research map / Image generation / Image restoration
Multitasks-based Deep Evidential Fusion Network for Blind Image Quality Assessment
Yiwei Lou, Yuanpeng He, Rongchao Zhang, Yongzhi Cao, Hanpin Wang, Yu Huang
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 scene and distortion classification auxiliary tasks (q to q+s+d) with all losses (ablation) | DEFNet · Six IQA datasets mean | SRCC: 0.922 -> 0.939 | Improved |
| add distortion-type task to quality-only model with L^M (ablation) | DEFNet · Six IQA datasets mean | SRCC: 0.910 -> 0.913 | Improved |
| add cross-region loss L^U (q+s+d) (ablation) | DEFNet · Six IQA datasets mean | SRCC: 0.916 -> 0.925 | Improved |
| add cross-grained loss L^F (q+s+d) (ablation) | DEFNet · Six IQA datasets mean | SRCC: 0.916 -> 0.926 | Improved |
| add both L^U and L^F (q+s+d) (ablation) | DEFNet · Six IQA datasets mean | PLCC: 0.906 -> 0.929 | Improved |
| add L^U and L^F to quality-only model (ablation) | DEFNet · Six IQA datasets mean | SRCC: 0.910 -> 0.922 | Improved |
| add L^U to q+s+d model (ablation) | DEFNet · Six IQA datasets mean | ACC_d: 0.851 -> 0.830 | Hurt |
| increase lambda_1 from 0.1 to 0.2 (lambda_2=0.1) (ablation) | DEFNet · LIVE | SRCC: 0.978 -> 0.981 | Improved |
| increase lambda_1 from 0.1 to 0.4 (lambda_2=0.1) (ablation) | DEFNet · LIVE-C | PLCC: 0.897 -> 0.821 | Hurt |
| increase lambda_2 from 0.1 to 0.3 (lambda_1=0.1) (ablation) | DEFNet · LIVE-C | PLCC: 0.897 -> 0.867 | Hurt |
| use DEFNet evidential fusion instead of LIQE | BIQA model · Mean confidence interval width | CI width: 0.286 -> 0.251 | Improved |
| use DEFNet instead of UNIQUE on TID2013 zero-shot | BIQA model · TID2013 (zero-shot) | SRCC: 0.768 -> 0.828 | Improved |
Closest papers by meaning
- Distilling Spatially-Heterogeneous Distortion Perception for Blind Image Quality Assessment (CVPR 2025 · similarity 0.85)
- Integrating Global Context Contrast and Local Sensitivity for Blind Image Quality Assessment (ICML 2024 · similarity 0.85)
- Feature Denoising Diffusion Model for Blind Image Quality Assessment (AAAI 2025 · similarity 0.83)
- Bridging the Synthetic-to-Authentic Gap: Distortion-Guided Unsupervised Domain Adaptation for Blind Image Quality Assessment (CVPR 2024 · similarity 0.83)
- Benchmarking Uncertainty Disentanglement: Specialized Uncertainties for Specialized Tasks (NeurIPS 2024 · similarity 0.83)
- Causal-IQA: Towards the Generalization of Image Quality Assessment Based on Causal Inference (ICML 2024 · similarity 0.83)
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.
