Diffeomorphic transforms for data augmentation of highly variable shape and texture objects.

Journal: Computer methods and programs in biomedicine
Published Date:

Abstract

BACKGROUND AND OBJECTIVE: Training a deep convolutional neural network (CNN) for automatic image classification requires a large database with images of labeled samples. However, in some applications such as biology and medicine only a few experts can correctly categorize each sample. Experts are able to identify small changes in shape and texture which go unnoticed by untrained people, as well as distinguish between objects in the same class that present drastically different shapes and textures. This means that currently available databases are too small and not suitable to train deep learning models from scratch. To deal with this problem, data augmentation techniques are commonly used to increase the dataset size. However, typical data augmentation methods introduce artifacts or apply distortions to the original image, which instead of creating new realistic samples, obtain basic spatial variations of the original ones.

Authors

  • Noelia Vállez
    VISILAB Group, ETSI Industriales, University of Castilla-La Mancha, Ciudad Real, Spain.
  • Gloria Bueno
    VISILAB Group, ETSI Industriales, University of Castilla-La Mancha, Ciudad Real, Spain.
  • Oscar Deniz
    VISILAB Group, ETSI Industriales, University of Castilla-La Mancha, Ciudad Real, Spain.
  • Saul Blanco
    Institute of the Environment, University of Leon, Leon E-24071, Spain.