# Two-stage Convolutional Part Heatmap Regression for the 1st 3D Face Alignment in the Wild (3DFAW) Challenge

arXiv: https://arxiv.org/abs/1609.09545 · HTML: https://mlsanity.com/paper/1609.09545/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Sep 2016.

## Built on by (most cited first, 8 of 8)

- [DAD-3DHeads: A Large-scale Dense, Accurate and Diverse Dataset for 3D Head Alignment from a Single Image](https://mlsanity.com/paper/2204.03688/) (2022)
- [SADRNet: Self-Aligned Dual Face Regression Networks for Robust 3D Dense Face Alignment and Reconstruction](https://mlsanity.com/paper/2106.03021/) (2021)
- [Pre-training strategies and datasets for facial representation learning](https://mlsanity.com/paper/2103.16554/) (2021)
- [Towards Robust and Unconstrained Full Range of Rotation Head Pose Estimation](https://mlsanity.com/paper/2309.07654/) (2023)
- [Subpixel Heatmap Regression for Facial Landmark Localization](https://mlsanity.com/paper/2111.02360/) (2021)
- [Expression-preserving face frontalization improves visually assisted speech processing](https://mlsanity.com/paper/2204.02810/) (2022)
- [A survey and classification of face alignment methods based on face models](https://mlsanity.com/paper/2311.03082/) (2023)
- [LDDMM-Face: Large Deformation Diffeomorphic Metric Learning for Flexible and Consistent Face Alignment](https://mlsanity.com/paper/2108.00690/) (2021)

## Abstract (arXiv)

This paper describes our submission to the 1st 3D Face Alignment in the Wild (3DFAW) Challenge. Our method builds upon the idea of convolutional part heatmap regression [1], extending it for 3D face alignment. Our method decomposes the problem into two parts: (a) X,Y (2D) estimation and (b) Z (depth) estimation. At the first stage, our method estimates the X,Y coordinates of the facial landmarks by producing a set of 2D heatmaps, one for each landmark, using convolutional part heatmap regression. Then, these heatmaps, alongside the input RGB image, are used as input to a very deep subnetwork trained via residual learning for regressing the Z coordinate. Our method ranked 1st in the 3DFAW Challenge, surpassing the second best result by more than 22%.