Visible to the public Style transfer-based image synthesis as an efficient regularization technique in deep learning

TitleStyle transfer-based image synthesis as an efficient regularization technique in deep learning
Publication TypeConference Paper
Year of Publication2019
AuthorsMikołajczyk, A., Grochowski, M.
Conference Name2019 24th International Conference on Methods and Models in Automation and Robotics (MMAR)
Date PublishedAug. 2019
ISBN Number978-1-7281-0933-6
Keywordsbase image, challenging skin lesion classification case study, convolutional neural nets, convolutional neural networks, data augmentation, Decision Support System, Deep Learning, deep neural networks, Diagnosis, efficient regularization technique, Image analysis, image classification, image classification purposes, image synthesis, labeled images, learning (artificial intelligence), learning algorithms, Lesions, machine learning, Metrics, Neural networks, neural style transfer, newly created images, pubcrawl, Regularization, regularization techniques e.g, relatively poor generalization abilities, representative neural architectures, resilience, Resiliency, Scalability, Skin, skin lesions, style transfer-based image synthesis, Task Analysis, Training, transfer learning, unlabeled images

These days deep learning is the fastest-growing area in the field of Machine Learning. Convolutional Neural Networks are currently the main tool used for the image analysis and classification purposes. Although great achievements and perspectives, deep neural networks and accompanying learning algorithms have some relevant challenges to tackle. In this paper, we have focused on the most frequently mentioned problem in the field of machine learning, that is relatively poor generalization abilities. Partial remedies for this are regularization techniques e.g. dropout, batch normalization, weight decay, transfer learning, early stopping and data augmentation. In this paper we have focused on data augmentation. We propose to use a method based on a neural style transfer, which allows to generate new unlabeled images of high perceptual quality that combine the content of a base image with the appearance of another one. In a proposed approach, the newly created images are described with pseudo-labels, and then used as a training dataset. Real, labeled images are divided into the validation and test set. We validated proposed method on a challenging skin lesion classification case study. Four representative neural architectures are examined. Obtained results show the strong potential of the proposed approach.

Citation Keymikolajczyk_style_2019