Research map / Image generation / Editing and style transfer
Unsupervised Creation of Parameterized Avatars
Lior Wolf, Yaniv Taigman, Adam Polyak
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 the tied output (configuration) terms of TOS instead of DTN, judged on e(..(x)) | TOS · Face emoji retrieval e(..(x)) | median rank: 18,079 -> 3,519 | Improved |
| train c jointly with G instead of fixing c to pretrained c-bar (ablation) | TOS · Face emoji retrieval e(..(x)) | median rank: 14,990 -> 3,519 | Improved |
| train c jointly with G instead of fixing c to pretrained c-bar (ablation) | TOS · VR avatars retrieval e(..(x)) | median rank: 43,160 -> 11,153 | Improved |
| use TOS instead of DTN on the raw generator output | TOS · Face emoji retrieval g(f(x)) | median rank: 16 -> 30 | Hurt |
| use TOS instead of DTN on the raw generator output | TOS · VR avatars retrieval g(f(x)) | median rank: 195 -> 758 | Improved |
| use TOS configuration instead of manually created emoji | TOS · Face emoji retrieval e(..(x)) | median rank: 16,311 -> 3,519 | Improved |
| use TOS instead of DANN domain adaptation baseline | TOS · VR avatars retrieval e(..(x)) | median rank: 52,435 -> 11,153 | Improved |
| use TOS with synthesis engine e instead of a plain DCGAN | TOS · Polygon toy problem | visual compliance with polygon class | Improved |
| apply pretrained c-bar on top of DTN output to obtain configuration | DTN · Face emoji retrieval e(..(x)) | median rank: 16 -> 18,079 | Hurt |
Closest papers by meaning
- Unsupervised Generation of Free-Form and Parameterized Avatars (TPAMI 2018 · similarity 0.97)
- Unsupervised Pixel-Level Domain Adaptation With Generative Adversarial Networks (CVPR 2017 · similarity 0.88)
- SwiftAvatar: Efficient Auto-Creation of Parameterized Stylized Character on Arbitrary Avatar Engines (AAAI 2023 · similarity 0.87)
- Unsupervised Cross-Domain Image Generation (ICLR 2017 · similarity 0.87)
- AvatarArtist: Open-Domain 4D Avatarization (CVPR 2025 · similarity 0.87)
- Cross-modal Latent Space Alignment for Image to Avatar Translation (ICCV 2023 · 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.
