Abstract:Deep neural networks have become the main work horse for many tasks involving learning from data in a variety of applications in Science and Engineering. Traditionally, the input to these networks lie in a vector space and the operations employed within the network are well defined on vector-spaces. In the recent past, due to technological advances in sensing, it has become possible to acquire manifold-valued data sets either directly or indirectly. Examples include but are not limited to data from omnidirectional cameras on automobiles, drones etc., synthetic aperture radar imaging, diffusion magnetic resonance imaging, elastography and conductance imaging in the Medical Imaging domain and others. Thus, there is need to generalize the deep neural networks to cope with input data that reside on curved manifolds where vector space operations are not naturally admissible. In this paper, we present a novel theoretical framework to generalize the widely popular convolutional neural networks (CNNs) to high dimensional manifold-valued data inputs. We call these networks, ManifoldNets. In ManifoldNets, convolution operation on data residing on Riemannian manifolds is achieved via a provably convergent recursive computation of the weighted Fréchet Mean (wFM) of the given data, where the weights makeup the convolution mask, to be learned. Further, we prove that the proposed wFM layer achieves a contraction mapping and hence ManifoldNet does not need the non-linear ReLU unit used in standard CNNs. We present experiments, using the ManifoldNet framework, to achieve dimensionality reduction by computing the principal linear subspaces that naturally reside on a Grassmannian. The experimental results demonstrate the efficacy of ManifoldNets in the context of classification and reconstruction accuracy.

Parallel Transport Convolution: A New Tool for Convolutional Neural Networks on Manifolds

Parallel Transport Convolution: Deformable Convolutional Networks on Manifold-Structured Data

NPTC-net: Narrow-Band Parallel Transport Convolutional Neural Network on Point Clouds

NPTC-net: Narrow-Band Parallel Transport Convolutional Neural Networks on Point Clouds

Dynamically Stable Poincaré Embeddings for Neural Manifolds

Deep Manifold Computing and Visualization

Manifold GCN: Diffusion-based Convolutional Neural Network for Manifold-valued Graphs

ManifoldNet: A Deep Network Framework for Manifold-valued Data

PFCNN: Convolutional Neural Networks on 3D Surfaces Using Parallel Frames

Learning Depthwise Separable Graph Convolution from Data Manifold

The Manifold Scattering Transform for High-Dimensional Point Cloud Data

PDE-based Group Equivariant Convolutional Neural Networks

A Tailored Convolutional Neural Network for Nonlinear Manifold Learning of Computational Physics Data using Unstructured Spatial Discretizations

A Tour of Convolutional Networks Guided by Linear Interpreters

A Convergence Rate for Manifold Neural Networks

Parallel Separable 3D Convolution for Video and Volumetric Data Understanding

Mumford-Shah and Potts Regularization for Manifold-Valued Data with Applications to DTI and Q-Ball Imaging

Convergence of Manifold Filter-Combine Networks

Exploring the Manifold of Neural Networks Using Diffusion Geometry

Geometric Wavelet Scattering Networks on Compact Riemannian Manifolds

Computational Optimal Transport and Filtering on Riemannian manifolds