Safe Augmentation: Learning Task-Specific Transformations from Data

Baran, Irynei

Home
→
Students Research & Project Works | Роботи студентів
→
Факультет прикладних наук
→
Освітня програма наук про дані
→
2019
→
View Item

dc.contributor.author	Baran, Irynei
dc.date.accessioned	2019-02-13T09:24:02Z
dc.date.available	2019-02-13T09:24:02Z
dc.date.issued	2019
dc.identifier.citation	Baran, Irynei. Safe Augmentation: Learning Task-Specific Transformations from Data : Master Thesis : manuscript / Irynei Baran ; Supervisor Arseny Kravchenko ; Ukrainian Catholic University, Department of Computer Sciences. – Lviv : [s.n.], 2019. – 23 p. : ill.	uk
dc.identifier.uri	http://er.ucu.edu.ua/handle/1/1312
dc.language.iso	en	uk
dc.subject	Data augmentation	uk
dc.subject	Deep learning	uk
dc.subject	Image classification	uk
dc.subject	Image segmentation	uk
dc.title	Safe Augmentation: Learning Task-Specific Transformations from Data	uk
dc.type	Preprint	uk
dc.status	Публікується вперше	uk
dc.description.abstracten	Data augmentation is widely used as a part of the training process applied to deep learning models, especially in the computer vision domain. Currently, common data augmentation techniques are designed manually. Therefore they require expert knowledge and time. Moreover, optimal augmentations found for one dataset, often do not transfer to other datasets as effectively. We propose a simple novel method that can automatically learn task-specific data augmentation techniques called safe augmentations that do not break the data distribution and can be used to improve model performance. Moreover, we provided a new training pipeline for using safe augmentations for different computer vision tasks. Our method works both with image classification and image segmentation and achieves significantly better accuracy on CIFAR-10, CIFAR-100, SVHN, Tiny ImageNet and Cityscapes datasets comparing to other augmentation techniques.	uk