Research map / Image generation / Autoregressive and latent generation
Attribute2Image: Conditional Image Generation from Visual Attributes
Xinchen Yan, Jimei Yang, Kihyuk Sohn, Honglak Lee
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 layered foreground/background modeling (disCVAE over CVAE) (ablation) | CVAE · LFW face reconstruction (full) | pixel MSE: 11.8 -> 10.0 | Improved |
| add layered modeling (ablation) | CVAE · LFW face completion: half image | pixel MSE: 20.3 -> 18.8 | Improved |
| add layered modeling (ablation) | CVAE · LFW face completion: eye region | pixel MSE: 12.0 -> 10.3 | Improved |
| add layered modeling (ablation) | CVAE · CUB bird reconstruction (full) | pixel MSE: 14.3 -> 12.9 | Improved |
| add layered modeling (ablation) | CVAE · CUB bird completion 8x8 | pixel MSE: 1.8 -> 1.8 | No clear effect |
| add layered modeling (ablation) | CVAE · CUB bird completion 16x16 | pixel MSE: 4.4 -> 4.4 | No clear effect |
| add attribute conditioning (CVAE over unconditional VAE) (ablation) | VAE · LFW face completion: face region | pixel MSE: 13.1 -> 12.3 | Improved |
| add attribute conditioning (ablation) | VAE · LFW face reconstruction (full) | pixel MSE: 11.8 -> 11.8 | No clear effect |
| add attribute conditioning (ablation) | VAE · CUB bird reconstruction (full) | pixel MSE: 14.5 -> 14.3 | Improved |
| add layered modeling (ablation) | CVAE · LFW attribute-conditioned generation | attribute MSE: 17.20 -> 16.71 | Improved |
| add layered modeling (ablation) | CVAE · LFW attribute-conditioned generation | cosine similarity: 0.9054 -> 0.9057 | No clear effect |
| generate from attributes instead of nearest-neighbor retrieval from training set | NN · LFW attribute-conditioned generation | attribute MSE: 21.88 -> 16.71 | Improved |
Closest papers by meaning
- Overcoming the Disentanglement vs Reconstruction Trade-off via Jacobian Supervision (ICLR 2019 · similarity 0.87)
- Unsupervised Robust Disentangling of Latent Characteristics for Image Synthesis (ICCV 2019 · similarity 0.87)
- Multi-View Data Generation Without View Supervision (ICLR 2018 · similarity 0.87)
- Explicit Disentanglement of Appearance and Perspective in Generative Models (NeurIPS 2019 · similarity 0.86)
- Boosting Generative Image Modeling via Joint Image-Feature Synthesis (NeurIPS 2025 · similarity 0.84)
- Attribute Based Interpretable Evaluation Metrics for Generative Models (ICML 2024 · similarity 0.84)
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.
