Content and organization section:

The success of deep learning in computer vision and image analysis, speech recognition, and natural language processing has driven the recent interest in developing similar models for 3D geometric data. However, it is less obvious how using convolutional neural networks (CNNs) architectures can be adapted to 3D data, given in the form of point clouds or meshes, where a regular structure is not directly available. The purpose of this course is to overview the foundations and the current state of the art in deep learning techniques for 3D shape analysis. This short course will cover the following topics:

– 3D human shape analysis.

– Basic Riemmannian geometry.

– Deep learning for 3D data: basic concepts of deep learning; extending CNN to 3D data;

– Generative methods for 3D data, autoencoders and GAN methods for 3D data.

The targeted applications will be in 3D face and 3D human body shape analysis.

Organizers: