Deep Signatures – Learning Invariants of Planar Curves

02/11/2022
by   Roy Velich, et al.
20

We propose a learning paradigm for numerical approximation of differential invariants of planar curves. Deep neural-networks' (DNNs) universal approximation properties are utilized to estimate geometric measures. The proposed framework is shown to be a preferable alternative to axiomatic constructions. Specifically, we show that DNNs can learn to overcome instabilities and sampling artifacts and produce numerically-stable signatures for curves subject to a given group of transformations in the plane. We compare the proposed schemes to alternative state-of-the-art axiomatic constructions of group invariant arc-lengths and curvatures.

READ FULL TEXT

page 8

page 9

research
03/06/2023

Learning Differential Invariants of Planar Curves

We propose a learning paradigm for the numerical approximation of differ...
research
11/23/2016

Learning Invariant Representations Of Planar Curves

We propose a metric learning framework for the construction of invariant...
research
03/30/2016

Möbius Invariants of Shapes and Images

Identifying when different images are of the same object despite changes...
research
02/03/2021

Length Learning for Planar Euclidean Curves

In this work, we used deep neural networks (DNNs) to solve a fundamental...
research
06/12/2008

Classification of curves in 2D and 3D via affine integral signatures

We propose a robust classification algorithm for curves in 2D and 3D, un...
research
10/11/2021

Using differential invariants to detect projective equivalences and symmetries of rational 3D curves

We present a new approach using differential invariants to detect projec...
research
07/03/2017

On Blockwise Symmetric Matchgate Signatures and Higher Domain #CSP

For any n≥ 3 and q≥ 3, we prove that the Equality function (=_n) on n ...

Please sign up or login with your details

Forgot password? Click here to reset