Data augmentation with Mobius transformations
This addresses the challenge of data scarcity in deep learning, offering a domain-specific improvement for image-based tasks.
The paper tackles the problem of improving deep model generalization, especially with scarce training data, by introducing a novel data augmentation method using Möbius transformations, which results in better performance compared to prior techniques like cutout and crop-and-flip.
Data augmentation has led to substantial improvements in the performance and generalization of deep models, and remain a highly adaptable method to evolving model architectures and varying amounts of data---in particular, extremely scarce amounts of available training data. In this paper, we present a novel method of applying Mobius transformations to augment input images during training. Mobius transformations are bijective conformal maps that generalize image translation to operate over complex inversion in pixel space. As a result, Mobius transformations can operate on the sample level and preserve data labels. We show that the inclusion of Mobius transformations during training enables improved generalization over prior sample-level data augmentation techniques such as cutout and standard crop-and-flip transformations, most notably in low data regimes.