Research map / 3D vision / Face and shape
A Decoupled 3D Facial Shape Model by Adversarial Training
Victoria Fernandez Abrevaya, Adnane Boukhayma, Stefanie Wuhrer, Edmond Boyer
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 mesh regularization (ablation) | Decoupled 3D face GAN · Decoupled 3D face model evaluation | specificity (mm): 3.61 -> 2.01 | Improved |
| add mesh regularization (ablation) | Decoupled 3D face GAN · Decoupled 3D face model evaluation | global diversity (mm): 1.41 -> 4.74 | Improved |
| add identity classification (ablation) | Decoupled 3D face GAN · Decoupled 3D face model evaluation | identity decoupling (%): 7.8 -> 98.6 | Improved |
| add expression classification (ablation) | Decoupled 3D face GAN · Decoupled 3D face model evaluation | expression decoupling (%): 42.8 -> 89.7 | Improved |
| add feature loss (ablation) | Decoupled 3D face GAN · Decoupled 3D face model evaluation | expression decoupling (%): 80.3 -> 89.7 | Improved |
| add feature loss (ablation) | Decoupled 3D face GAN · Decoupled 3D face model evaluation | global diversity (mm): 4.47 -> 4.74 | Improved |
| add feature loss (ablation) | Decoupled 3D face GAN · Decoupled 3D face model evaluation | identity decoupling (%): 96.0 -> 98.6 | Hurt |
| use adversarial 3D-2D model instead of 3DMM | Decoupled 3D face GAN · Decoupled 3D face model evaluation | expression decoupling (%): 65.6 -> 89.7 | Improved |
| use adversarial 3D-2D model instead of MAE autoencoder | Decoupled 3D face GAN · Decoupled 3D face model evaluation | expression decoupling (%): 53.3 -> 89.7 | Improved |
| use adversarial 3D-2D model instead of CoMA/DeepFLAME | Decoupled 3D face GAN · Decoupled 3D face model evaluation | global diversity (mm): 3.38 -> 4.74 | Improved |
| use adversarial 3D-2D model instead of 3DMM | Decoupled 3D face GAN · Decoupled 3D face model evaluation | identity decoupling (%): 99.6 -> 98.6 | Hurt |
| use adversarial 3D-2D model instead of CoMA | Decoupled 3D face GAN · Sparse landmark reconstruction (lambda=0) | RMSE with no regularization (mm): 3.05 -> 2.62 | Improved |
Closest papers by meaning
- 3DFaceGAN: Adversarial Nets for 3D Face Representation, Generation, and Translation (IJCV 2020 · similarity 0.91)
- Unpaired Multi-domain Attribute Translation of 3D Facial Shapes with a Square and Symmetric Geometric Map (ICCV 2023 · similarity 0.91)
- Generating 3D Faces using Convolutional Mesh Autoencoders (ECCV 2018 · similarity 0.90)
- Synthesizing Coupled 3D Face Modalities by Trunk-Branch Generative Adversarial Networks (ECCV 2020 · similarity 0.90)
- GANFIT: Generative Adversarial Network Fitting for High Fidelity 3D Face Reconstruction (CVPR 2019 · similarity 0.90)
- A 3D GAN for Improved Large-Pose Facial Recognition (CVPR 2021 · similarity 0.89)
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.
