Offset equivariant networks and their applications

Marco Cotogni, Claudio Cusano
7/1/2022
cs.CV

No official code found. Best implementation: claudio-unipv/offset-equivariant (6 stars, MIT, last commit 2023-10-01).

Abstract

In this paper we present a framework for the design and implementation of offset equivariant networks, that is, neural networks that preserve in their output uniform increments in the input. In a suitable color space this kind of networks achieves equivariance with respect to the photometric transformations that characterize changes in the lighting conditions. We verified the framework on three different problems: image recognition, illuminant estimation, and image inpainting. Our experiments show that the performance of offset equivariant networks are comparable to those in the state of the art on regular data. Differently from conventional networks, however, equivariant networks do behave consistently well when the color of the illuminant changes.

View on arXivView PDF

Code Implementations(1)

claudio-unipv/offset-equivariantCommunity100%
MIT

Offset equivariant networks

61PythonApr 20, 20223 years agoMIT

Similar Papers

A Hypertoroidal Covering for Perfect Color Equivariance

88% match2026

Learning Color Equivariant Representations

86% match1 repo12024

Learning Color Equivariant Representations

86% match1 repo12024

Color Equivariant Convolutional Networks

82% match1 repo102023

Latent Equivariant Operators for Robust Object Recognition: Promise and Challenges

81% match1 repo22026

Equivariant Denoisers for Image Restoration

81% match2024

Cite this paper

@article{cotogni2022offset,
  title  = {Offset equivariant networks and their applications},
  author = {Marco Cotogni and Claudio Cusano},
  year   = {2022},
  eprint = {2207.00292},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2207.00292v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2207.00292.svg)](https://codeofpaper.com/paper/2207.00292)

Discussion