Research map / Image generation / Text-to-image diffusion
Personalized Image Aesthetics
Jian Ren, Xiaohui Shen, Zhe Lin, Radomir Mech, David J. Foran
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 |
|---|---|---|---|
| jointly train attributes and aesthetics instead of attributes alone | attribute prediction network · AADB attribute prediction (VC) | Spearman correlation: 0.5759 -> 0.6938 | Improved |
| jointly train attributes and aesthetics instead of attributes alone | attribute prediction network · AADB attribute prediction (RoT) | Spearman correlation: 0.2174 -> 0.2737 | Improved |
| use residual-based adaptation (PAM) instead of direct SVM score prediction | personalized aesthetics model · FLICKR-AES (10 images) | correlation improvement: -0.352 -> 0.006 | Improved |
| use residual-based adaptation (PAM) instead of direct SVM score prediction | personalized aesthetics model · FLICKR-AES (100 images) | correlation improvement: -0.176 -> 0.039 | Improved |
| use PAM instead of feature-based matrix factorization (FPMF) | personalized aesthetics model · FLICKR-AES (100 images) | correlation improvement: 0.010 -> 0.039 | Improved |
| add content features to attribute features (ablation) | PAM · FLICKR-AES (100 images) | correlation improvement: 0.025 -> 0.039 | Improved |
| add attribute features to content features (ablation) | PAM · FLICKR-AES (100 images) | correlation improvement: 0.021 -> 0.039 | Improved |
| add content features to attribute features (ablation) | PAM · FLICKR-AES (10 images) | correlation improvement: 0.004 -> 0.006 | Improved |
| add more user training images (10 to 100) | PAM · FLICKR-AES | correlation improvement: 0.006 -> 0.039 | Improved |
| add more user training images (10 to 100) | FPMF (content and attribute) · FLICKR-AES | correlation improvement: -0.001 -> 0.010 | No clear effect |
| use Active-PAM selection instead of random selection | PAM · FLICKR-AES and REAL-CUR | average correlation | Improved |
| use Active-PAM selection instead of Greedy, MCAL, QBC | PAM · FLICKR-AES and REAL-CUR (unseen images) | correlation improvement | Improved |
Closest papers by meaning
- Correct for Whom? Subjectivity and the Evaluation of Personalized Image Aesthetics Assessment Models (AAAI 2023 · similarity 0.92)
- Scaling Up Personalized Image Aesthetic Assessment via Task Vector Customization (ECCV 2024 · similarity 0.91)
- Learning Personalized Photographic Style from Pairwise User Preferences (CVPR 2026 · similarity 0.91)
- Advancing Comprehensive Aesthetic Insight with Multi-Scale Text-Guided Self-Supervised Learning (AAAI 2025 · similarity 0.88)
- Aligning Vision Models with Human Aesthetics in Retrieval: Benchmarks and Algorithms (NeurIPS 2024 · similarity 0.87)
- Aesthetic Post-Training Diffusion Models from Generic Preferences with Step-by-step Preference Optimization (CVPR 2025 · 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.
