Research map / Vision-language and multimodal / CLIP and zero-shot
HDR-VLM: HDR-Domain Adaptation of VLMs and Preference-Aligned Quality Assessment for HDR Video Color Grading
Hao Yuan, Jiabin Zhang, Yajing Wu, Ruixuan Pang, Jing Li
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 Stage 2 GRPO preference alignment after Stage 1 domain adaptation (ablation) | HDR-VLM · HDR color grading dataset | PLCC: 0.8024 -> 0.9033 | Improved |
| add Stage 1 HDR domain adaptation before RL (ablation) | HDR-VLM · HDR color grading dataset | PLCC: 0.8140 -> 0.9033 | Improved |
| add Stage 1 SFT domain adaptation to a pretrained VLM | Qwen2.5-VL 7B · HDR color grading dataset | PLCC: 0.7828 -> 0.8024 | Improved |
| use progressive unfreezing instead of one-shot full-model fine-tuning in Stage 1 (ablation) | HDR domain adaptation · HDR domain-adaptation test set | accuracy: 0.85 -> 0.9 | Improved |
| unfreeze projection MLP jointly with ViT (ablation) | HDR domain adaptation · HDR domain-adaptation test set | accuracy: 0.79 -> 0.82 | Improved |
| tune the ViT only on HDR data (ablation) | pre-trained VLM · HDR domain-adaptation test set | accuracy: 0.54 -> 0.79 | Improved |
| add Step 2 exogenous rewards (RL2 + Rdist) after Step 1 ranking rewards (ablation) | HDR-VLM GRPO · HDR color grading dataset | PLCC: 0.8860 -> 0.9033 | Improved |
| add Step 2 exogenous rewards after Step 1 (ablation) | HDR-VLM GRPO · HDR color grading dataset | SROCC: 0.8659 -> 0.8667 | No clear effect |
| add Step 1 endogenous ranking rewards before Step 2 rewards (ablation) | HDR-VLM GRPO · HDR color grading dataset | SROCC: 0.7703 -> 0.8667 | Improved |
| use HDR-VLM progressive reward instead of VQAThinker reward | VQAThinker · HDR color grading dataset | PLCC: 0.8745 -> 0.9033 | Improved |
Closest papers by meaning
- Perceptual Assessment and Optimization of HDR Image Rendering (CVPR 2024 · similarity 0.88)
- KVQ: Boosting Video Quality Assessment via Saliency-guided Local Perception (CVPR 2025 · similarity 0.86)
- Q-Insight: Understanding Image Quality via Visual Reinforcement Learning (NeurIPS 2025 · similarity 0.85)
- MegaCoin: Enhancing Medium-Grained Color Perception for Vision-Language Models (AAAI 2026 · similarity 0.85)
- Q-Bench: A Benchmark for General-Purpose Foundation Models on Low-level Vision (ICLR 2024 · similarity 0.85)
- White-Balance First, Adjust Later: Cross-Camera Color Constancy via Vision-Language Evaluation (CVPR 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.
