Research map / Image generation / Super-resolution
Attention-Aware Face Hallucination via Deep Reinforcement Learning
Check 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 |
|---|---|---|---|
| use sequential attention-guided patch enhancement rather than GLN | Face hallucination · LFW-funneled 4× | PSNR: 30.34 -> 32.93 | Helped |
| use sequential attention-guided patch enhancement rather than GLN | Face hallucination · LFW-funneled 8× | PSNR: 24.51 -> 26.17 | Helped |
| use sequential attention-guided patch enhancement rather than GLN | Face hallucination · BioID 4× | PSNR: 29.13 -> 31.56 | Helped |
| use sequential attention-guided patch enhancement rather than GLN | Face hallucination · BioID 8× | PSNR: 24.76 -> 26.56 | Helped |
| increase the number of enhancement steps from 5 to 25 (ablation) | Attention-FH · LFW 4× | PSNR: 28.89 -> 32.93 | Helped |
| increase the number of enhancement steps from 5 to 25 (ablation) | Attention-FH · LFW 8× | PSNR: 23.55 -> 26.17 | Helped |
| increase the number of enhancement steps from 25 to 35 (ablation) | Attention-FH · LFW 4× | PSNR: 32.93 -> 32.91 | No clear effect |
| increase the number of enhancement steps from 25 to 35 (ablation) | Attention-FH · LFW 8× | PSNR: 26.17 -> 26.31 | Helped |
| add local patch attention instead of recurrently enhancing the whole image (ablation) | Attention-FH · LFW 4× | PSNR: 32.26 -> 32.93 | Helped |
| add local patch attention instead of recurrently enhancing the whole image (ablation) | Attention-FH · LFW 8× | PSNR: 25.71 -> 26.17 | Helped |
| use policy-selected patches instead of random patches (ablation) | Attention-FH · LFW 4× | PSNR: 31.60 -> 32.93 | Helped |
| use policy-selected patches instead of random patches (ablation) | Attention-FH · LFW 8× | PSNR: 25.76 -> 26.17 | Helped |
Closest papers by meaning
- Dynamic Multimodal Activation Steering for Hallucination Mitigation in Large Vision-Language Models (ICLR 2026 · similarity 0.87)
- Locate-then-Sparsify: Attribution Guided Sparse Strategy for Visual Hallucination Mitigation (CVPR 2026 · similarity 0.87)
- Causally-Grounded Dual-Path Attention Intervention for Object Hallucination Mitigation in LVLMs (AAAI 2026 · similarity 0.87)
- Mitigating Object Hallucinations in Large Vision-Language Models with Assembly of Global and Local Attention (CVPR 2025 · similarity 0.87)
- Ground What You See: Hallucination-Resistant MLLMs via Caption Feedback, Diversity-Aware Sampling, and Conflict Regularization (AAAI 2026 · similarity 0.86)
- P$^2$-DPO:Grounding Hallucination in Perceptual Processing via Calibration Direct Preference Optimization (ICLR 2026 · 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.