Data Augmentation to Improve the Performance of a Convolutional Neural Network on Image Classification

Deshan Fonseka; Christos Chrysoulas

doi:10.1109/DASA51403.2020.9317249

Data Augmentation to Improve the Performance of a Convolutional Neural Network on Image Classification

Deshan Fonseka, Christos Chrysoulas

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

6 Citations (Scopus)

Abstract

Deep learning has become a fundamental tool to extract meaningful information from big data. However, it needs a huge amount of high-quality data to build an accurate classifier. In many situations, the size of the training dataset is not sufficiently large to effectively train a model. This paper presents a Convolutional Neural Network trained on a very small dataset, discussing the impact of data augmentation, feature extraction and fine-tuning on the accuracy of the model. The results show that having a small dataset, those approaches are very effective when dealing with image data.</xpl-document-abstract>

Original language	English
Title of host publication	2020 International Conference on Decision Aid Sciences and Application (DASA)
Publisher	IEEE
ISBN (Electronic)	9781728196770
ISBN (Print)	9781728196787
DOIs	https://doi.org/10.1109/DASA51403.2020.9317249
Publication status	Published - 15 Jan 2021
Event	International Conference on Decision Aid Sciences and Application 2020 - Sakheer, Bahrain Duration: 7 Nov 2020 → 9 Nov 2020 https://dasa20.uob.edu.bh/

Conference

Conference	International Conference on Decision Aid Sciences and Application 2020
Abbreviated title	DASA 2020
Country/Territory	Bahrain
City	Sakheer
Period	7/11/20 → 9/11/20
Internet address	https://dasa20.uob.edu.bh/

Keywords

Convolutional Neural Network
Data Augmentation
Deep Learning
Transfer Learning

ASJC Scopus subject areas

Artificial Intelligence
Computer Science Applications
Decision Sciences (miscellaneous)
Information Systems and Management
Control and Optimization

Access to Document

10.1109/DASA51403.2020.9317249

Cite this

@inproceedings{9087dee6fdd44dc2a621905288563844,

title = "Data Augmentation to Improve the Performance of a Convolutional Neural Network on Image Classification",

abstract = "Deep learning has become a fundamental tool to extract meaningful information from big data. However, it needs a huge amount of high-quality data to build an accurate classifier. In many situations, the size of the training dataset is not sufficiently large to effectively train a model. This paper presents a Convolutional Neural Network trained on a very small dataset, discussing the impact of data augmentation, feature extraction and fine-tuning on the accuracy of the model. The results show that having a small dataset, those approaches are very effective when dealing with image data.",

keywords = "Convolutional Neural Network, Data Augmentation, Deep Learning, Transfer Learning",

author = "Deshan Fonseka and Christos Chrysoulas",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; International Conference on Decision Aid Sciences and Application 2020, DASA 2020 ; Conference date: 07-11-2020 Through 09-11-2020",

year = "2021",

month = jan,

day = "15",

doi = "10.1109/DASA51403.2020.9317249",

language = "English",

isbn = "9781728196787",

booktitle = "2020 International Conference on Decision Aid Sciences and Application (DASA)",

publisher = "IEEE",

address = "United States",

url = "https://dasa20.uob.edu.bh/",

}

Fonseka, D & Chrysoulas, C 2021, Data Augmentation to Improve the Performance of a Convolutional Neural Network on Image Classification. in 2020 International Conference on Decision Aid Sciences and Application (DASA)., 9317249, IEEE, International Conference on Decision Aid Sciences and Application 2020, Sakheer, Bahrain, 7/11/20. https://doi.org/10.1109/DASA51403.2020.9317249

TY - GEN

T1 - Data Augmentation to Improve the Performance of a Convolutional Neural Network on Image Classification

AU - Fonseka, Deshan

AU - Chrysoulas, Christos

PY - 2021/1/15

Y1 - 2021/1/15

N2 - Deep learning has become a fundamental tool to extract meaningful information from big data. However, it needs a huge amount of high-quality data to build an accurate classifier. In many situations, the size of the training dataset is not sufficiently large to effectively train a model. This paper presents a Convolutional Neural Network trained on a very small dataset, discussing the impact of data augmentation, feature extraction and fine-tuning on the accuracy of the model. The results show that having a small dataset, those approaches are very effective when dealing with image data.

AB - Deep learning has become a fundamental tool to extract meaningful information from big data. However, it needs a huge amount of high-quality data to build an accurate classifier. In many situations, the size of the training dataset is not sufficiently large to effectively train a model. This paper presents a Convolutional Neural Network trained on a very small dataset, discussing the impact of data augmentation, feature extraction and fine-tuning on the accuracy of the model. The results show that having a small dataset, those approaches are very effective when dealing with image data.

KW - Convolutional Neural Network

KW - Data Augmentation

KW - Deep Learning

KW - Transfer Learning

UR - http://www.scopus.com/inward/record.url?scp=85100567791&partnerID=8YFLogxK

U2 - 10.1109/DASA51403.2020.9317249

DO - 10.1109/DASA51403.2020.9317249

M3 - Conference contribution

AN - SCOPUS:85100567791

SN - 9781728196787

BT - 2020 International Conference on Decision Aid Sciences and Application (DASA)

PB - IEEE

T2 - International Conference on Decision Aid Sciences and Application 2020

Y2 - 7 November 2020 through 9 November 2020

ER -

Data Augmentation to Improve the Performance of a Convolutional Neural Network on Image Classification

Abstract

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this