Conditional Invertible Flow for Point Cloud Generation

Stypułkowski, Michał; Zamorski, Maciej; Zięba, Maciej; Chorowski, Jan

doi:10.48550/arxiv.1910.07344

Cited by 4 publications

(5 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Volumetric GANs [Wu et al 2016] can create 3D models with a voxel grid representation. Normalizing flows [Dinh et al 2015;Rezende and Mohamed 2015] (NF) have been extended to 3D by modeling the distribution of point clouds as an invertible parameterized transformation from a probability density embedded in 3D space [Kim et al 2020;Stypułkowski et al 2019;]. In the following, we mainly discuss variational autoencoders (VAEs) and auto-regressive models (ARs) for 3D tasks that are closely related to our roof synthesis application.…”

Section: Generative Modelingmentioning

confidence: 99%

Intuitive and efficient roof modeling for reconstruction and synthesis

Ren

Zhang

et al. 2021

ACM Trans. Graph.

View full text Add to dashboard Cite

We propose a novel and flexible roof modeling approach that can be used for constructing planar 3D polygon roof meshes. Our method uses a graph structure to encode roof topology and enforces the roof validity by optimizing a simple but effective planarity metric we propose. This approach is significantly more efficient than using general purpose 3D modeling tools such as 3ds Max or SketchUp, and more powerful and expressive than specialized tools such as the straight skeleton. Our optimization-based formulation is also flexible and can accommodate different styles and user preferences for roof modeling. We showcase two applications. The first application is an interactive roof editing framework that can be used for roof design or roof reconstruction from aerial images. We highlight the efficiency and generality of our approach by constructing a mesh-image paired dataset consisting of 2539 roofs. Our second application is a generative model to synthesize new roof meshes from scratch. We use our novel dataset to combine machine learning and our roof optimization techniques, by using transformers and graph convolutional networks to model roof topology, and our roof optimization methods to enforce the planarity constraint.

show abstract

Section: Generative Modelingmentioning

confidence: 99%

Intuitive and efficient roof modeling for reconstruction and synthesis

Ren

Zhang

et al. 2021

ACM Trans. Graph.

View full text Add to dashboard Cite

show abstract

“…Moreover, ShapeGF [5] applies an energy-based framework to model a shape by learning the gradient field of its log-density and recover the distributions of the points. [45] models point clouds using conditional NFs. However, they do not learn the low-dimensional shape embeddings leading to poor performance compared to related work.…”

Section: Generative Models For Point Cloudsmentioning

confidence: 99%

Go with the Flows: Mixtures of Normalizing Flows for Point Cloud Generation and Reconstruction

Postels¹,

Liu²,

Spezialetti³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…In such a framework, we use an additional continuous normalizing flow G ψ , which transfers latent space into a Gaussian prior. Finally, we propose to use a decoder that returns weights of the target network D θ : Z z → Θ, instead of 3D points as done in [14,15]. The resulting hypernetwork contains an encoder E φ , a decoder D θ and a flow G ψ (Fig.…”

Section: Airplanementioning

confidence: 99%

“…More recent methods that create representations of 3D objects from variable-size unordered point clouds rely on generative neural networks that treat point clouds as a sample from a 3D probability distribution [14,15,16]. PointFlow [14] returns probability distributions of the 3D object point cloud, instead of an exact set of points.…”

Section: Introductionmentioning

confidence: 99%

HyperFlow: Representing 3D Objects as Surfaces

Spurek¹,

Zięba²,

Tabor³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

In this work, we present HyperFlow -a novel generative model that leverages hypernetworks to create continuous 3D object representations in a form of lightweight surfaces (meshes), directly out of point clouds. Efficient object representations are essential for many computer vision applications, including robotic manipulation and autonomous driving. However, creating those representations is often cumbersome, because it requires processing unordered sets of point clouds. Therefore, it is either computationally expensive, due to additional optimization constraints such as permutation invariance, or leads to quantization losses introduced by binning point clouds into discrete voxels. Inspired by mesh-based representations of objects used in computer graphics, we postulate a fundamentally different approach and represent 3D objects as a family of surfaces. To that end, we devise a generative model that uses a hypernetwork to return the weights of a Continuous Normalizing Flows (CNF) target network. The goal of this target network is to map points from a probability distribution into a 3D mesh. To avoid numerical instability of the CNF on compact support distributions, we propose a new Spherical Log-Normal function which models density of 3D points around object surfaces mimicking noise introduced by 3D capturing devices. As a result, we obtain continuous meshbased object representations that yield better qualitative results than competing approaches, while reducing training time by over an order of magnitude.

show abstract

Conditional Invertible Flow for Point Cloud Generation

Cited by 4 publications

References 0 publications

Intuitive and efficient roof modeling for reconstruction and synthesis

Intuitive and efficient roof modeling for reconstruction and synthesis

Go with the Flows: Mixtures of Normalizing Flows for Point Cloud Generation and Reconstruction

HyperFlow: Representing 3D Objects as Surfaces

Contact Info

Product

Resources

About