3D Face Reconstruction with the Geometric Guidance of Facial Part Segmentation

TL;DR

The paper addresses accurate 3D face reconstruction under extreme expressions by leveraging facial part segmentation through Part Re-projection Distance Loss (PRDL). By transforming segmentation into 2D point sets and using grid anchors with multiple distance statistics, PRDL provides gradient-rich geometry guidance that improves alignment between reconstructed facial parts and the input image, outperforming differentiable silhouette renderers. It introduces new mesh-part annotations aligned with 2D segmentation definitions and a synthetic emotional-expression dataset to bolster training. Across Part IoU and REALY benchmarks, PRDL achieves state-of-the-art overlap and 3D accuracy, enabling more faithful and expressive 3D reconstructions for VR/AR and CGI applications.

Abstract

3D Morphable Models (3DMMs) provide promising 3D face reconstructions in various applications. However, existing methods struggle to reconstruct faces with extreme expressions due to deficiencies in supervisory signals, such as sparse or inaccurate landmarks. Segmentation information contains effective geometric contexts for face reconstruction. Certain attempts intuitively depend on differentiable renderers to compare the rendered silhouettes of reconstruction with segmentation, which is prone to issues like local optima and gradient instability. In this paper, we fully utilize the facial part segmentation geometry by introducing Part Re-projection Distance Loss (PRDL). Specifically, PRDL transforms facial part segmentation into 2D points and re-projects the reconstruction onto the image plane. Subsequently, by introducing grid anchors and computing different statistical distances from these anchors to the point sets, PRDL establishes geometry descriptors to optimize the distribution of the point sets for face reconstruction. PRDL exhibits a clear gradient compared to the renderer-based methods and presents state-of-the-art reconstruction performance in extensive quantitative and qualitative experiments. Our project is available at https://github.com/wang-zidu/3DDFA-V3 .

Figures (19)

• Figure 1: We introduce Part Re-projection Distance Loss (PRDL) for 3D face reconstruction, leveraging the geometric guidance provided by facial part segmentation. PRDL enhances the alignment of reconstructed facial features with the original image and excels in capturing extreme expressions.
• Figure 2: Drawbacks of existing research and our results. (a) Present researches fail to reconstruct extreme expressions and perform bad region alignment. (b) Inconsistencies between 3D errors and 2D alignments, such as the eye region in this case. (c) Geometric optimization of each semantically consistent part is only achievable through our annotations.
• Figure 3: Overview of Part Re-projection Distance Loss (PRDL). (a): Transforming facial part segmentation into target point sets $\{{\bm{C}_p}\}$. (b): Re-projecting ${V_{3d}}(\bm{\alpha} )$ onto the image plane to obtain predicted point sets $\{ {V_{2d}^p(\bm{\alpha} )\}}$. (c): Given anchors $\bm{A}$ and distance functions $\bm{\mathcal{F}}$, the core idea of PRDL is to minimize the difference of every statistical distance from any $\bm{a}_i \in \bm{A}$ to the ${V_{2d}^p(\bm{\alpha} )}$ or $\bm{C}_p$, leading to enhanced overlap between the regions covered by the target and predicted point sets.
• Figure 4: Synthesize emotional expression data.
• Figure 5: Examples of our synthetic face dataset.