Self‐supervised learning with randomised layers for remote sensing

This letter presents a new self‐supervised learning approach based on randomised layers for remote sensing. Our method is basically based on the Tile2Vec approach, which is one of the state‐of‐the‐art self‐supervised learning approaches for remote sensing. Unlike the original Tile2Vec algorithm, we...

Full description

Saved in:
Bibliographic Details
Published in:Electronics letters Vol. 57; no. 6; pp. 249 - 251
Main Authors: Jung, Heechul, Jeon, Taegyun
Format: Journal Article
Language:English
Published: Stevenage John Wiley & Sons, Inc 01.03.2021
Wiley
Subjects:
Algorithms
Classification
Combinatorial mathematics
Computer vision and image processing techniques
Datasets
Image recognition
Instrumentation and techniques for geophysical, hydrospheric and lower atmosphere research
Machine learning
Neural networks
Programming and algorithm theory
Regression analysis
Remote sensing
Supervised learning
ISSN:0013-5194, 1350-911X
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This letter presents a new self‐supervised learning approach based on randomised layers for remote sensing. Our method is basically based on the Tile2Vec approach, which is one of the state‐of‐the‐art self‐supervised learning approaches for remote sensing. Unlike the original Tile2Vec algorithm, we reformulate the triplet loss as a classification loss. We use several fully connected layers with binary cross‐entropy loss instead of no fully connected layers with triplet loss of the original Tile2Vec. We observe that not updating the fully connected layers is more helpful in obtaining more robust representations. The proposed algorithm is verified and evaluated by applying it to a cropland data layer classification task. The experimental results show that our approach is superior to the original Tile2Vec approach in all experiments based on random forest, logistic regression, and multi‐layer classifiers.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0013-5194
1350-911X
DOI:10.1049/ell2.12108