Beltrami Signature: A Novel Invariant 2D Shape Representation for Object Classification

03/30/2021
by   Chenran Lin, et al.
0

There is a growing interest in shape analysis in recent years and in this paper we present a novel contour-based shape representation named Beltrami signature for 2D bounded simple connected domain. The proposed representation is based on conformal welding. With suitable normalization, the uniqueness of welding is guaranteed up to a rotation. Then it can be extended to a harmonic function and finally quasi-conformal theory get rid of the only uncertainty by computing Beltrami coefficient of harmonic extension. The benifits of the proposed signature is it keeps invariant under simple transformations like sacling, transformation and rotation and is roubost under slight deformation and distortion. Experiments demonstrates the above properties and also shows the excellent classification performance.

READ FULL TEXT

page 7

page 9

page 21

page 22

page 23

page 24

research
10/19/2010

ANSIG - An Analytic Signature for Arbitrary 2D Shapes (or Bags of Unlabeled Points)

In image analysis, many tasks require representing two-dimensional (2D) ...
research
08/15/2019

Shape Optimization of Rotating Electric Machines using Isogeometric Analysis and Harmonic Stator-Rotor Coupling

This work deals with shape optimization of electric machines using isoge...
research
04/12/2018

CubeNet: Equivariance to 3D Rotation and Translation

3D Convolutional Neural Networks are sensitive to transformations applie...
research
06/27/2019

Effective Rotation-invariant Point CNN with Spherical Harmonics kernels

We present a novel rotation invariant architecture operating directly on...
research
04/17/2022

Learning with Signatures

In this work we investigate the use of the Signature Transform in the co...
research
03/21/2022

Disentangling Patterns and Transformations from One Sequence of Images with Shape-invariant Lie Group Transformer

An effective way to model the complex real world is to view the world as...

Please sign up or login with your details

Forgot password? Click here to reset