Research map / Image generation / Editing and style transfer
Assessing Eye Aesthetics for Automatic Multi-Reference Eye In-Painting
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 aesthetic assessment to select the reference instead of SSIM selection (ablation) | AesGAN · Celeb-ID test | MeanL1: 4.82E-3 -> 4.76E-3 | Helped |
| add parsing loss (ablation) | AesGAN with aesthetic reference selection · Celeb-ID test | IS: 4.04 -> 4.09 | Helped |
| add parsing loss (ablation) | AesGAN with aesthetic reference selection · Celeb-ID test | FID: 6.99 -> 6.55 | Helped |
| add aesthetic loss using one-branch AesNet (ablation) | AesGAN baseline · Celeb-ID test | MeanL1: 4.82E-3 -> 4.74E-3 | Helped |
| add aesthetic loss using two-branch AesNet (ablation) | AesGAN baseline · Celeb-ID test | FID: 6.74 -> 6.78 | Hurt |
| add parsing loss and aesthetic loss using one-branch AesNet (ablation) | AesGAN · Celeb-ID test | MeanL1: 4.78E-3 -> 4.67E-3 | Helped |
| add parsing loss and aesthetic loss using two-branch AesNet (ablation) | AesGAN · Celeb-ID test | IS: 4.11 -> 4.15 | Helped |
| add residual blocks to the aesthetic assessment model (ablation) | AesNet · Eye aesthetic dataset | accuracy: 0.7 -> 0.73 | Helped |
| add a reconstruction branch to AesNet with residual blocks (ablation) | AesNet · Eye aesthetic dataset | accuracy: 0.73 -> 0.84 | Helped |
| use aesthetic assessment guidance and losses rather than ExGAN's random exemplar selection | AesGAN · Celeb-ID test | FID: 15.66 -> 6.43 | Helped |
| use aesthetic assessment guided in-painting rather than multi-reference MR-GAN | AesGAN · Celeb-ID test | MeanL1: 11.77E-3 -> 4.68E-3 | Helped |
| apply eye aesthetic constraints during in-painting | AesGAN · Celeb-ID test | eye aesthetic category: 10.1% of the test images changed from low quality to high quality | Helped |
Closest papers by meaning
- Bridging Cognitive Gap: Hierarchical Description Learning for Artistic Image Aesthetics Assessment (AAAI 2026 · similarity 0.86)
- Unlocking the Essence of Beauty: Advanced Aesthetic Reasoning with Relative-Absolute Policy Optimization (ICLR 2026 · similarity 0.85)
- Advancing Comprehensive Aesthetic Insight with Multi-Scale Text-Guided Self-Supervised Learning (AAAI 2025 · similarity 0.84)
- Can Vision–Language Models Assess Graphic Design Aesthetics? A Benchmark, Evaluation, and Dataset Perspective. (ICLR 2026 · similarity 0.83)
- ArtiMuse: Fine-Grained Image Aesthetics Assessment with Joint Scoring and Expert-Level Understanding (CVPR 2026 · similarity 0.83)
- Venus: Benchmarking and Empowering Multimodal Large Language Models for Aesthetic Guidance and Cropping (CVPR 2026 · 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.