Scopus İndeksli Yayınlar Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12416/8651
Browse
Search Results
Article Distribution-Preserving Data Augmentation(PeerJ Inc., 2021) Saran, Nurdan Ayse; Nar, Fatih; Saran, MuratConference Object Citation - WoS: 9Citation - Scopus: 16Perlin Random Erasing for Data Augmentation(Ieee, 2021) Saran, Ayse Nurdan; Saran, Murat; Nar, FatihIn the last decade, Deep Learning is applied in a wide range of problems with tremendous success. Large data, increased computational resources, and theoretical improvements are main reasons for this success. As the dataset grows, the real-world is better represented, allows developing a model that can generalize. However, creating a labeled dataset is expensive, time-consuming, or sometimes even challenging. Therefore, researchers proposed data augmentation methods to increase dataset size by creating variations of the existing data. This study proposes an extension to Random Erasing data augmentation method by introducing smoothness. The proposed method provides better performance compared to Random Erasing data augmentation method, which is shown using a transfer learning scenario on the UC Merced Land-use image dataset.Article Citation - WoS: 4Citation - Scopus: 6Distribution-Preserving Data Augmentation(Peerj inc, 2021) Nar, Fatih; Saran, Nurdan Ayse; Saran, MuratIn the last decade, deep learning has been applied in a wide range of problems with tremendous success. This success mainly comes from large data availability, increased computational power, and theoretical improvements in the training phase. As the dataset grows, the real world is better represented, making it possible to develop a model that can generalize. However, creating a labeled dataset is expensive, time-consuming, and sometimes not likely in some domains if not challenging. Therefore, researchers proposed data augmentation methods to increase dataset size and variety by creating variations of the existing data. For image data, variations can be obtained by applying color or spatial transformations, only one or a combination. Such color transformations perform some linear or nonlinear operations in the entire image or in the patches to create variations of the original image. The current color-based augmentation methods are usually based on image processing methods that apply color transformations such as equalizing, solarizing, and posterizing. Nevertheless, these color-based data augmentation methods do not guarantee to create plausible variations of the image. This paper proposes a novel distribution-preserving data augmentation method that creates plausible image variations by shifting pixel colors to another point in the image color distribution. We achieved this by defining a regularized density decreasing direction to create paths from the original pixels' color to the distribution tails. The proposed method provides superior performance compared to existing data augmentation methods which is shown using a transfer learning scenario on the UC Merced Land-use, Intel Image Classification, and Oxford-IIIT Pet datasets for classification and segmentation tasks.