Generating 3D faces using Convolutional Mesh Autoencoders

Ranjan, Anurag; Bolkart, Timo; Sanyal, Soubhik; Black, Michael J.

doi:10.48550/arxiv.1807.10267

Cited by 7 publications

(11 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The authors used convolutional operators from [64] and showed examples of human body shape completion from partial scans. A follow-up work CoMA [57] used a similar architecture with spectral Chebyshev filters [22] and additional spatial pooling to generate 3D facial meshes. The authors claim that CoMA can represent better faces with expressions than PCA in a very small dimensional latent space of only eight dimensions.…”

Section: Geometric Deep Learningmentioning

confidence: 99%

“…Fully connected layers have huge number of parameters and also do not take into account the local geometric of the 3D facial surfaces. The only method that represented faces using convolutions on the mesh domain was the recently proposed mesh auto-encoder CoMA [57]. Nevertheless, the identity and expression latent space of CoMA was mixed.…”

Section: D Facial Shape Representation and Generationmentioning

confidence: 99%

“…The mesh downsampling step is performed by the surface simplification method in [25], which minimises the quadric error when decimating the template. Up-sampling of the template is based on the barycentric coordinates of the contracted vertices in decimated mesh [57]. In total, we perform 4 levels of down-sampling, with each level lowering the number of vertices by approximately 4 times.…”

Section: Meshgan Architecturementioning

confidence: 99%

“…Intrinsic generative models are currently a key open question in geometric deep learning. Intrinsic auto-encoder architectures have recently been proposed for human body [45] and face [57] meshes. Nevertheless, due to the lack of appropriate adversarial training, these auto-encoders retain only the lowpass shape information and lose most of the details.…”

Section: Introductionmentioning

confidence: 99%

“…• We present the first GAN architecture for 3D face generation. Contrary to the auto-encoder recently proposed in [57] that learns latent spaces where identity and expression are mixed, we can generate expression for arbitrary identities.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

MeshGAN: Non-linear 3D Morphable Models of Faces

Cheng¹,

Bronstein²,

Zhou³

et al. 2019

Preprint

View full text Add to dashboard Cite

Generative Adversarial Networks (GANs) are currently the method of choice for generating visual data. Certain GAN architectures and training methods have demonstrated exceptional performance in generating realistic synthetic images (in particular, of human faces). However, for 3D object, GANs still fall short of the success they have had with images. One of the reasons is due to the fact that so far GANs have been applied as 3D convolutional architectures to discrete volumetric representations of 3D objects. In this paper, we propose the first intrinsic GANs architecture operating directly on 3D meshes (named as MeshGAN). Both quantitative and qualitative results are provided to show that MeshGAN can be used to generate high-fidelity 3D face with rich identities and expressions.

show abstract

Section: Geometric Deep Learningmentioning

confidence: 99%

Section: D Facial Shape Representation and Generationmentioning

confidence: 99%

Section: Meshgan Architecturementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

MeshGAN: Non-linear 3D Morphable Models of Faces

Cheng¹,

Bronstein²,

Zhou³

et al. 2019

Preprint

View full text Add to dashboard Cite

show abstract

High-Quality Face Capture Using Anatomical Muscles

Bao

Cong

Grabli

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

Muscle-based systems have the potential to provide both anatomical accuracy and semantic interpretability as compared to blendshape models; however, a lack of expressivity and differentiability has limited their impact. Thus, we propose modifying a recently developed rather expressive muscle-based system in order to make it fully-differentiable; in fact, our proposed modifications allow this physically robust and anatomically accurate muscle model to conveniently be driven by an underlying blendshape basis. Our formulation is intuitive, natural, as well as monolithically and fully coupled such that one can differentiate the model from end to end, which makes it viable for both optimization and learning-based approaches for a variety of applications. We illustrate this with a number of examples including both shape matching of three-dimensional geometry as as well as the automatic determination of a threedimensional facial pose from a single two-dimensional RGB image without using markers or depth information.

show abstract

Combining 3D Morphable Models: A Large Scale Face-And-Head Model

Ploumpis

Wang

Pears

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

framework to blend covariance matrices from multiple models. Second, as an example application of our approach, we build a new face-and-head shape model that combines the variability and facial detail of the LSFM with the full head modelling of the LYHM. The resulting combined shape model achieves state-of-the-art performance and outperforms existing head models by a large margin. Finally, as an application experiment, we reconstruct full head representations from single, unconstrained images by utilizing our proposed large-scale model in conjunction with the Face-Warehouse blendshapes for handling expressions.

show abstract

Generating 3D faces using Convolutional Mesh Autoencoders

Cited by 7 publications

References 0 publications

MeshGAN: Non-linear 3D Morphable Models of Faces

MeshGAN: Non-linear 3D Morphable Models of Faces

High-Quality Face Capture Using Anatomical Muscles

Combining 3D Morphable Models: A Large Scale Face-And-Head Model

Contact Info

Product

Resources

About