Research map / Image generation / Editing and style transfer
Supervised Attribute Information Removal and Reconstruction for Image Manipulation
Nannan Li, Bryan A. Plummer
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 disentanglement loss (attribute remover) (ablation) | AIRR · CelebA-HQ | Avg. manipulation accuracy: 64.5 -> 73.7 | Improved |
| add disentanglement loss (attribute remover) (ablation) | AIRR · DeepFashion Fine-grained Attribute | Avg. manipulation accuracy: 60.8 -> 62.1 | Improved |
| add disentanglement loss (attribute remover) (ablation) | AIRR · DeepFashion Synthesis | Avg. manipulation accuracy: 91.7 -> 95.3 | Improved |
| add disentanglement loss (attribute remover) (ablation) | AIRR · CelebA | Avg. manipulation accuracy: 81.8 -> 86.2 | Improved |
| add parsing mask and background encoder branch (ablation) | AIRR · DeepFashion Fine-grained Attribute | Avg. manipulation accuracy: 45.4 -> 62.1 | Improved |
| add parsing mask and background encoder branch (ablation) | AIRR · DeepFashion Synthesis | R@5: 94.4 -> 97.6 | Improved |
| add parsing mask and background encoder branch (ablation) | AIRR · CelebA | Avg. manipulation accuracy: 83.0 -> 86.2 | Improved |
| use disentanglement loss instead of honesty loss (ablation) | AIRR · DeepFashion Fine-grained Attribute | R@5: 51.1 -> 57.7 | Improved |
| use disentanglement loss instead of honesty loss (ablation) | AIRR · DeepFashion Synthesis | Avg. manipulation accuracy: 90.0 -> 95.3 | Improved |
| AIRR vs AttGAN, strongest baseline | AIRR · CelebA | Avg. manipulation accuracy: 69.5 -> 86.2 | Improved |
| AIRR vs InterfaceGAN with ablated model (w/o mask) | AIRR (w/o mask) · CelebA-HQ | R@5: 19.1 -> 60.9 | Improved |
| user study: AIRR vs StarGAN | AIRR · DeepFashion Fine-grained Attribute | user preference: 65%/35% | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- InterfaceGAN — InterFaceGAN: Interpreting the Disentangled Face Representation Learned by GANs (TPAMI 2020) · better on 2 of 2 shared results
- StarGAN — StarGAN: Unified Generative Adversarial Networks for Multi-Domain Image-to-Image Translation (CVPR 2018) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Memory-Augmented Attribute Manipulation Networks for Interactive Fashion Search (CVPR 2017 · similarity 0.84)
- Learning Residual Images for Face Attribute Manipulation (CVPR 2017 · similarity 0.84)
- Any Fashion Attribute Editing: Dataset and Pretrained Models (TPAMI 2025 · similarity 0.84)
- Learning Attribute-Driven Disentangled Representations for Interactive Fashion Retrieval (ICCV 2021 · similarity 0.83)
- Gradual Residuals Alignment: A Dual-Stream Framework for GAN Inversion and Image Attribute Editing (AAAI 2024 · similarity 0.82)
- SDGAN: Disentangling Semantic Manipulation for Facial Attribute Editing (AAAI 2024 · similarity 0.82)
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.
