Research map / Vision-language and multimodal / Human and embodied modeling
Blind Image Quality Assessment via Vision-Language Correspondence: A Multitask Learning Perspective
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 scene classification and distortion type identification to BIQA training (ablation) | LIQE · Six IQA datasets | SRCC: 0.915 -> 0.922 | Helped |
| Add distortion type identification to BIQA training (ablation) | LIQE · Six IQA datasets | SRCC: 0.915 -> 0.920 | Helped |
| Add scene classification alone to BIQA training (ablation) | LIQE · Six IQA datasets | SRCC: 0.915 -> 0.915 | No clear effect |
| Fine-tune pretrained CLIP on IQA data rather than using it pretrained (ablation) | CLIP BIQA · Six IQA datasets | Mean SRCC: 0.693 -> 0.890 | Helped |
| Use five quality levels instead of two (ablation) | LIQE · Six IQA datasets | Mean SRCC: 0.890 -> 0.922 | Helped |
| Fine-tune rather than freeze the textual encoder (ablation) | LIQE · Six IQA datasets | Mean SRCC: 0.918 -> 0.922 | Helped |
| Use one joint probability formulation rather than separate task templates (ablation) | LIQE · Six IQA datasets | Mean SRCC: 0.916 -> 0.922 | Helped |
| Use dynamic loss weighting rather than equal task weights (ablation) | LIQE · Six IQA datasets | Mean SRCC: 0.920 -> 0.922 | Helped |
| Train with multitask vision-language correspondence rather than visual-encoder-only fine-tuning (ablation) | LIQE · Six IQA datasets | Mean SRCC: 0.915 -> 0.922 | Helped |
| Use LIQE rather than UNIQUE for BIQA | BIQA models · TID2013 | SRCC: 0.768 -> 0.811 | Helped |
| Use LIQE rather than UNIQUE for BIQA | BIQA models · SPAQ | SRCC: 0.838 -> 0.881 | Helped |
| Use LIQE rather than UNIQUE for BIQA | BIQA models · PIPAL | SRCC: 0.444 -> 0.478 | Helped |
Closest papers by meaning
- Integrating Global Context Contrast and Local Sensitivity for Blind Image Quality Assessment (ICML 2024 · similarity 0.92)
- Distilling Spatially-Heterogeneous Distortion Perception for Blind Image Quality Assessment (CVPR 2025 · similarity 0.90)
- Semi-Supervised Blind Image Quality Assessment through Knowledge Distillation and Incremental Learning (AAAI 2024 · similarity 0.89)
- Feature Denoising Diffusion Model for Blind Image Quality Assessment (AAAI 2025 · similarity 0.89)
- Q-Bench: A Benchmark for General-Purpose Foundation Models on Low-level Vision (ICLR 2024 · similarity 0.88)
- Bridging the Synthetic-to-Authentic Gap: Distortion-Guided Unsupervised Domain Adaptation for Blind Image Quality Assessment (CVPR 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.