Quantifying The Generative Capabilities Of Variational Autoencoders For 3D Car Point Clouds

Sneha Saha1; Stefan Menzel1; Leandro L. Minku2; Xin Yao2,3; Bernhard Sendhoff1; Patricia Wollstadt1

10.1109/SSCI47803.2020.9308513

2020

2020 IEEE Symposium Series on Computational Intelligence (SSCI)

978-1-7281-2548-0

2020 IEEE SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE (SSCI)

1469-1477

2020-12-01

Canberra, ACT, Australia

345 E 47TH ST, NEW YORK, NY 10017 USA

IEEE

During each cycle of automotive development, large amounts of geometric data are generated as results of design studies and simulation tasks. Discovering hidden knowledge from this data and making it available to the development team strengthens the design process by utilizing historic information when creating novel products. To this end, we propose to use powerful geometric deep learning models that learn low-dimensional representation of the design data in an unsupervised fashion. Trained models allow to efficiently explore the design space, as well as to generate novel designs. One popular class of generative models are variational autoencoders, which have however been rarely applied to geometric data. Hence, we use a variational autoencoder for 3D point clouds (PC-VAE) and explore the model's generative capabilities with a focus on the generation of realistic yet novel 3D shapes. We apply the PC-VAE to point clouds sampled from car shapes from a benchmark data set and employ quantitative measures to show that our PC-VAE generates realistic car shapes, wile returning a richer variety of unseen shapes compared to a baseline autoencoder. Finally, we demonstrate how the PC-VAE can be guided towards generating shapes with desired target properties by optimizing the parameters that maximize the output of a trained classifier for said target properties. We conclude that generative models are a powerful tool that may aid designers in automotive product development.

Representation learning geometric deep learning point clouds generative model novelty

其他

英语

CPCI-S

European Union's Horizon 2020 research and innovation programme[766186]

Computer Science ; Engineering

Computer Science, Artificial Intelligence ; Computer Science, Interdisciplinary Applications ; Engineering, Electrical & Electronic

WOS:000682772901067

Web of Science

1.Honda Research Institute Europe GmbH, Carl-Legien-Str. 30, Offenbach, Germany
2.CERCIA, School of Computer Science, University of Birmingham, Birmingham, UK
3.Department of Computer Science and Engineering, Southern University of Science and Technology, Shenzhen, China

Sneha Saha,Stefan Menzel,Leandro L. Minku,et al. Quantifying The Generative Capabilities Of Variational Autoencoders For 3D Car Point Clouds[C]. 345 E 47TH ST, NEW YORK, NY 10017 USA:IEEE,2020:1469-1477.