Robust Training of Social Media Image Classification Models

<p>Images shared on social media help crisis managers gain situational awareness and assess incurred damages, among other response tasks. As the volume and velocity of such content are typically high, real-time image classification has become an urgent need for faster disaster response. Recent...

Full description

Saved in:
Bibliographic Details
Main Author: Firoj Alam (14158866) (author)
Other Authors: Tanvirul Alam (14150628) (author), Ferda Ofli (8983517) (author), Muhammad Imran (282621) (author)
Published: 2022
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1864513562726105088
author Firoj Alam (14158866)
author2 Tanvirul Alam (14150628)
Ferda Ofli (8983517)
Muhammad Imran (282621)
author2_role author
author
author
author_facet Firoj Alam (14158866)
Tanvirul Alam (14150628)
Ferda Ofli (8983517)
Muhammad Imran (282621)
author_role author
dc.creator.none.fl_str_mv Firoj Alam (14158866)
Tanvirul Alam (14150628)
Ferda Ofli (8983517)
Muhammad Imran (282621)
dc.date.none.fl_str_mv 2022-12-26T00:00:00Z
dc.identifier.none.fl_str_mv 10.1109/tcss.2022.3230839
dc.relation.none.fl_str_mv https://figshare.com/articles/journal_contribution/Robust_Training_of_Social_Media_Image_Classification_Models/24056223
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Information and computing sciences
Computer vision and multimedia computation
Data management and data science
Machine learning
Task analysis
Social networking (online)
Data models
Training
Benchmark testing
Real-time systems
Pipelines
Crisis informatics
Disaster response
Humanitarian tasks
Multitask learning
Social media image classification
dc.title.none.fl_str_mv Robust Training of Social Media Image Classification Models
dc.type.none.fl_str_mv Text
Journal contribution
info:eu-repo/semantics/publishedVersion
text
contribution to journal
description <p>Images shared on social media help crisis managers gain situational awareness and assess incurred damages, among other response tasks. As the volume and velocity of such content are typically high, real-time image classification has become an urgent need for faster disaster response. Recent advances in computer vision and deep neural networks have enabled the development of models for image classification for a number of tasks, including detecting crisis incidents, filtering irrelevant images, classifying images into specific humanitarian categories, and assessing the severity of the damage. To develop robust models, it is necessary to understand the capability of the publicly available pretrained models for these tasks, which remains to be underexplored in the crisis informatics literature. In this study, we address such limitations by investigating ten different network architectures for four different tasks using the largest publicly available datasets for these tasks. We also explore various data augmentation strategies, semisupervised techniques, and a multitask learning setup. In our extensive experiments, we achieve promising results.</p><h2>Other Information</h2><p>Published in: IEEE Transactions on Computational Social Systems<br>License: <a href="https://creativecommons.org/licenses/by/4.0/legalcode" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/tcss.2022.3230839" target="_blank">https://dx.doi.org/10.1109/tcss.2022.3230839</a></p>
eu_rights_str_mv openAccess
id Manara2_b87d57deba056d769ef6b98d40935d53
identifier_str_mv 10.1109/tcss.2022.3230839
network_acronym_str Manara2
network_name_str Manara2
oai_identifier_str oai:figshare.com:article/24056223
publishDate 2022
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Robust Training of Social Media Image Classification ModelsFiroj Alam (14158866)Tanvirul Alam (14150628)Ferda Ofli (8983517)Muhammad Imran (282621)Information and computing sciencesComputer vision and multimedia computationData management and data scienceMachine learningTask analysisSocial networking (online)Data modelsTrainingBenchmark testingReal-time systemsPipelinesCrisis informaticsDisaster responseHumanitarian tasksMultitask learningSocial media image classification<p>Images shared on social media help crisis managers gain situational awareness and assess incurred damages, among other response tasks. As the volume and velocity of such content are typically high, real-time image classification has become an urgent need for faster disaster response. Recent advances in computer vision and deep neural networks have enabled the development of models for image classification for a number of tasks, including detecting crisis incidents, filtering irrelevant images, classifying images into specific humanitarian categories, and assessing the severity of the damage. To develop robust models, it is necessary to understand the capability of the publicly available pretrained models for these tasks, which remains to be underexplored in the crisis informatics literature. In this study, we address such limitations by investigating ten different network architectures for four different tasks using the largest publicly available datasets for these tasks. We also explore various data augmentation strategies, semisupervised techniques, and a multitask learning setup. In our extensive experiments, we achieve promising results.</p><h2>Other Information</h2><p>Published in: IEEE Transactions on Computational Social Systems<br>License: <a href="https://creativecommons.org/licenses/by/4.0/legalcode" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1109/tcss.2022.3230839" target="_blank">https://dx.doi.org/10.1109/tcss.2022.3230839</a></p>2022-12-26T00:00:00ZTextJournal contributioninfo:eu-repo/semantics/publishedVersiontextcontribution to journal10.1109/tcss.2022.3230839https://figshare.com/articles/journal_contribution/Robust_Training_of_Social_Media_Image_Classification_Models/24056223CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/240562232022-12-26T00:00:00Z
spellingShingle Robust Training of Social Media Image Classification Models
Firoj Alam (14158866)
Information and computing sciences
Computer vision and multimedia computation
Data management and data science
Machine learning
Task analysis
Social networking (online)
Data models
Training
Benchmark testing
Real-time systems
Pipelines
Crisis informatics
Disaster response
Humanitarian tasks
Multitask learning
Social media image classification
status_str publishedVersion
title Robust Training of Social Media Image Classification Models
title_full Robust Training of Social Media Image Classification Models
title_fullStr Robust Training of Social Media Image Classification Models
title_full_unstemmed Robust Training of Social Media Image Classification Models
title_short Robust Training of Social Media Image Classification Models
title_sort Robust Training of Social Media Image Classification Models
topic Information and computing sciences
Computer vision and multimedia computation
Data management and data science
Machine learning
Task analysis
Social networking (online)
Data models
Training
Benchmark testing
Real-time systems
Pipelines
Crisis informatics
Disaster response
Humanitarian tasks
Multitask learning
Social media image classification