Research map / 3D vision / Face and shape
“Look Ma, no landmarks!” – Unsupervised, Model-based Dense Face Alignment
Tatsuro Koizumi, William A. P. Smith
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 camera intrinsics regularisation Eint (ablation) | Ours · Stirling/ESRC 3D Face Database (HQ) | error: 2.65 -> 2.39 | Improved |
| add camera intrinsics regularisation Eint (ablation) | Ours · Stirling/ESRC 3D Face Database (LQ) | error: 2.60 -> 2.55 | Improved |
| add camera intrinsics regularisation Eint (ablation) | Ours · Stirling/ESRC 3D Face Database (Full) | error: 2.62 -> 2.49 | Improved |
| add robust residual loss Eres on top of reconstruction loss and statistical regularisation (ablation) | Ours · CelebA qualitative reconstructions | shrinking of reconstructed face | Improved |
| evaluate landmarks from fitted model instead of direct correspondence | Ours · AFLW (Mean[0-90]) | NME: 16.00 -> 18.63 | Hurt |
| evaluate landmarks from fitted model instead of direct correspondence | Ours · AFLW2000-3D (Mean[0-90]) | NME: 16.63 -> 18.55 | Hurt |
| evaluate landmarks from fitted model instead of direct correspondence | Ours · AFLW (Mean[0-30]) | NME: 5.51 -> 5.87 | Hurt |
| evaluate landmarks from fitted model instead of direct correspondence | Ours · AFLW2000-3D (Mean[0-30]) | NME: 4.98 -> 4.74 | Mixed |
| aggregate multiple frames in a single optimisation over pixel-wise predictions | Ours · multiframe qualitative (Fig. 7) | reconstruction quality | Improved |
Closest papers by meaning
- Learning Dense Facial Correspondences in Unconstrained Images (ICCV 2017 · similarity 0.91)
- Unsupervised Training for 3D Morphable Model Regression (CVPR 2018 · similarity 0.89)
- Joint 3D Face Reconstruction and Dense Alignment with Position Map Regression Network (ECCV 2018 · similarity 0.88)
- DSFNet: Dual Space Fusion Network for Occlusion-Robust 3D Dense Face Alignment (CVPR 2023 · similarity 0.88)
- MMFace: A Multi-Metric Regression Network for Unconstrained Face Reconstruction (CVPR 2019 · similarity 0.88)
- Faster Than Real-Time Facial Alignment: A 3D Spatial Transformer Network Approach in Unconstrained Poses (ICCV 2017 · similarity 0.88)
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.
