Comparison of Different Image Data Augmentation Approaches
Convolutional neural networks (CNNs) have gained prominence in the research literature on image classification over the last decade. One shortcoming of CNNs, however, is their lack of generalizability and tendency to overfit when presented with small training sets. Augmentation directly confronts th...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-11-01
|
| Series: | Journal of Imaging |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2313-433X/7/12/254 |
| _version_ | 1797503354023182336 |
|---|---|
| author | Loris Nanni Michelangelo Paci Sheryl Brahnam Alessandra Lumini |
| author_facet | Loris Nanni Michelangelo Paci Sheryl Brahnam Alessandra Lumini |
| author_sort | Loris Nanni |
| collection | DOAJ |
| description | Convolutional neural networks (CNNs) have gained prominence in the research literature on image classification over the last decade. One shortcoming of CNNs, however, is their lack of generalizability and tendency to overfit when presented with small training sets. Augmentation directly confronts this problem by generating new data points providing additional information. In this paper, we investigate the performance of more than ten different sets of data augmentation methods, with two novel approaches proposed here: one based on the discrete wavelet transform and the other on the constant-Q Gabor transform. Pretrained ResNet50 networks are finetuned on each augmentation method. Combinations of these networks are evaluated and compared across four benchmark data sets of images representing diverse problems and collected by instruments that capture information at different scales: a virus data set, a bark data set, a portrait dataset, and a LIGO glitches data set. Experiments demonstrate the superiority of this approach. The best ensemble proposed in this work achieves state-of-the-art (or comparable) performance across all four data sets. This result shows that varying data augmentation is a feasible way for building an ensemble of classifiers for image classification. |
| first_indexed | 2024-03-10T03:49:24Z |
| format | Article |
| id | doaj.art-09c130e3da5d4a5991b3aee8555bdbb1 |
| institution | Directory Open Access Journal |
| issn | 2313-433X |
| language | English |
| last_indexed | 2024-03-10T03:49:24Z |
| publishDate | 2021-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Journal of Imaging |
| spelling | doaj.art-09c130e3da5d4a5991b3aee8555bdbb12023-11-23T09:00:33ZengMDPI AGJournal of Imaging2313-433X2021-11-0171225410.3390/jimaging7120254Comparison of Different Image Data Augmentation ApproachesLoris Nanni0Michelangelo Paci1Sheryl Brahnam2Alessandra Lumini3Department of Information Engineering, University of Padua, Via Gradenigo 6, 35131 Padova, ItalyBioMediTech, Faculty of Medicine and Health Technology, Tampere University, Arvo Ylpön katu 34, FI-33520 Tampere, FinlandComputer Information Systems, Missouri State University, 901 S. National, Springfield, MO 65804, USADipartimento di Informatica–Scienza e Ingegneria (DISI), Università di Bologna, Via dell’Università 50, 47521 Cesena, ItalyConvolutional neural networks (CNNs) have gained prominence in the research literature on image classification over the last decade. One shortcoming of CNNs, however, is their lack of generalizability and tendency to overfit when presented with small training sets. Augmentation directly confronts this problem by generating new data points providing additional information. In this paper, we investigate the performance of more than ten different sets of data augmentation methods, with two novel approaches proposed here: one based on the discrete wavelet transform and the other on the constant-Q Gabor transform. Pretrained ResNet50 networks are finetuned on each augmentation method. Combinations of these networks are evaluated and compared across four benchmark data sets of images representing diverse problems and collected by instruments that capture information at different scales: a virus data set, a bark data set, a portrait dataset, and a LIGO glitches data set. Experiments demonstrate the superiority of this approach. The best ensemble proposed in this work achieves state-of-the-art (or comparable) performance across all four data sets. This result shows that varying data augmentation is a feasible way for building an ensemble of classifiers for image classification.https://www.mdpi.com/2313-433X/7/12/254data augmentationdeep learningconvolutional neural networksensemble |
| spellingShingle | Loris Nanni Michelangelo Paci Sheryl Brahnam Alessandra Lumini Comparison of Different Image Data Augmentation Approaches Journal of Imaging data augmentation deep learning convolutional neural networks ensemble |
| title | Comparison of Different Image Data Augmentation Approaches |
| title_full | Comparison of Different Image Data Augmentation Approaches |
| title_fullStr | Comparison of Different Image Data Augmentation Approaches |
| title_full_unstemmed | Comparison of Different Image Data Augmentation Approaches |
| title_short | Comparison of Different Image Data Augmentation Approaches |
| title_sort | comparison of different image data augmentation approaches |
| topic | data augmentation deep learning convolutional neural networks ensemble |
| url | https://www.mdpi.com/2313-433X/7/12/254 |
| work_keys_str_mv | AT lorisnanni comparisonofdifferentimagedataaugmentationapproaches AT michelangelopaci comparisonofdifferentimagedataaugmentationapproaches AT sherylbrahnam comparisonofdifferentimagedataaugmentationapproaches AT alessandralumini comparisonofdifferentimagedataaugmentationapproaches |