Convolutional neural network of isoplane polyhedron spherical gauge

A method of generating a convolutional neural network for operating on a spherical manifold generates locally defined gauges at a plurality of locations on the spherical manifold. The convolution is defined at each location on the spherical manifold with reference to an arbitrarily selected local de...

Full description

Saved in:
Bibliographic Details
Main Authors KIKKANAGLU BARIS, COHEN, THOMAS, S, DE HAAN PETER
Format Patent
LanguageChinese
English
Published 29.07.2022
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:A method of generating a convolutional neural network for operating on a spherical manifold generates locally defined gauges at a plurality of locations on the spherical manifold. The convolution is defined at each location on the spherical manifold with reference to an arbitrarily selected local definition gauge. The result of a convolution defined based on gauge isovariance at each location is transformed to obtain a manifold convolution. 一种生成用于在球面流形上操作的卷积神经网络的方法在该球面流形上的多个位置处生成局部定义的量规。在球面流形上的每个位置处参照任意选择的局部定义量规来定义卷积。在每个位置处基于量规等变性来定义的卷积的结果被转化以获得流形卷积。
Bibliography:Application Number: CN202080065492