A Comparative Study of Knowledge Transfer Methods for Misaligned Urban Building Labels ...
Misalignment in Earth observation (EO) images and building labels impact the training of accurate convolutional neural networks (CNNs) for semantic segmentation of building footprints. Recently, three Teacher-Student knowledge transfer methods have been introduced to address this issue: supervised d...
Main Authors: | , , |
---|---|
Format: | Report |
Language: | unknown |
Published: |
arXiv
2023
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.2311.03867 https://arxiv.org/abs/2311.03867 |
id |
ftdatacite:10.48550/arxiv.2311.03867 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.2311.03867 2023-12-31T10:06:17+01:00 A Comparative Study of Knowledge Transfer Methods for Misaligned Urban Building Labels ... Neupane, Bipul Aryal, Jagannath Rajabifard, Abbas 2023 https://dx.doi.org/10.48550/arxiv.2311.03867 https://arxiv.org/abs/2311.03867 unknown arXiv Creative Commons Attribution Non Commercial No Derivatives 4.0 International https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode cc-by-nc-nd-4.0 Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences CreativeWork Preprint article Article 2023 ftdatacite https://doi.org/10.48550/arxiv.2311.03867 2023-12-01T10:44:38Z Misalignment in Earth observation (EO) images and building labels impact the training of accurate convolutional neural networks (CNNs) for semantic segmentation of building footprints. Recently, three Teacher-Student knowledge transfer methods have been introduced to address this issue: supervised domain adaptation (SDA), knowledge distillation (KD), and deep mutual learning (DML). However, these methods are merely studied for different urban buildings (low-rise, mid-rise, high-rise, and skyscrapers), where misalignment increases with building height and spatial resolution. In this study, we present a workflow for the systematic comparative study of the three methods. The workflow first identifies the best (with the highest evaluation scores) hyperparameters, lightweight CNNs for the Student (among 43 CNNs from Computer Vision), and encoder-decoder networks (EDNs) for both Teachers and Students. Secondly, three building footprint datasets are developed to train and evaluate the identified Teachers and ... : This work has been submitted to Elsevier for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible ... Report DML DataCite Metadata Store (German National Library of Science and Technology) |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
unknown |
topic |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences |
spellingShingle |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences Neupane, Bipul Aryal, Jagannath Rajabifard, Abbas A Comparative Study of Knowledge Transfer Methods for Misaligned Urban Building Labels ... |
topic_facet |
Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences |
description |
Misalignment in Earth observation (EO) images and building labels impact the training of accurate convolutional neural networks (CNNs) for semantic segmentation of building footprints. Recently, three Teacher-Student knowledge transfer methods have been introduced to address this issue: supervised domain adaptation (SDA), knowledge distillation (KD), and deep mutual learning (DML). However, these methods are merely studied for different urban buildings (low-rise, mid-rise, high-rise, and skyscrapers), where misalignment increases with building height and spatial resolution. In this study, we present a workflow for the systematic comparative study of the three methods. The workflow first identifies the best (with the highest evaluation scores) hyperparameters, lightweight CNNs for the Student (among 43 CNNs from Computer Vision), and encoder-decoder networks (EDNs) for both Teachers and Students. Secondly, three building footprint datasets are developed to train and evaluate the identified Teachers and ... : This work has been submitted to Elsevier for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible ... |
format |
Report |
author |
Neupane, Bipul Aryal, Jagannath Rajabifard, Abbas |
author_facet |
Neupane, Bipul Aryal, Jagannath Rajabifard, Abbas |
author_sort |
Neupane, Bipul |
title |
A Comparative Study of Knowledge Transfer Methods for Misaligned Urban Building Labels ... |
title_short |
A Comparative Study of Knowledge Transfer Methods for Misaligned Urban Building Labels ... |
title_full |
A Comparative Study of Knowledge Transfer Methods for Misaligned Urban Building Labels ... |
title_fullStr |
A Comparative Study of Knowledge Transfer Methods for Misaligned Urban Building Labels ... |
title_full_unstemmed |
A Comparative Study of Knowledge Transfer Methods for Misaligned Urban Building Labels ... |
title_sort |
comparative study of knowledge transfer methods for misaligned urban building labels ... |
publisher |
arXiv |
publishDate |
2023 |
url |
https://dx.doi.org/10.48550/arxiv.2311.03867 https://arxiv.org/abs/2311.03867 |
genre |
DML |
genre_facet |
DML |
op_rights |
Creative Commons Attribution Non Commercial No Derivatives 4.0 International https://creativecommons.org/licenses/by-nc-nd/4.0/legalcode cc-by-nc-nd-4.0 |
op_doi |
https://doi.org/10.48550/arxiv.2311.03867 |
_version_ |
1786838275566075904 |