Dreaming More Data: Class-dependent Distributions over Diffeomorphisms for Learned Data Augmentation
This work addresses the need for automated, end-to-end learning in data augmentation for computer vision, though it is incremental as it builds on existing augmentation concepts.
The paper tackled the problem of manual specification in data augmentation by learning class-specific transformations within a diffeomorphism framework, resulting in significant performance improvements in training deep neural networks over traditional methods.
Data augmentation is a key element in training high-dimensional models. In this approach, one synthesizes new observations by applying pre-specified transformations to the original training data; e.g.~new images are formed by rotating old ones. Current augmentation schemes, however, rely on manual specification of the applied transformations, making data augmentation an implicit form of feature engineering. With an eye towards true end-to-end learning, we suggest learning the applied transformations on a per-class basis. Particularly, we align image pairs within each class under the assumption that the spatial transformation between images belongs to a large class of diffeomorphisms. We then learn a class-specific probabilistic generative models of the transformations in a Riemannian submanifold of the Lie group of diffeomorphisms. We demonstrate significant performance improvements in training deep neural nets over manually-specified augmentation schemes. Our code and augmented datasets are available online.