Trajectory-aware Principal Manifold Framework for Data Augmentation and Image Generation
This work addresses data augmentation and image generation for deep learning applications, offering a novel approach to manifold-based sample generation, though it appears incremental in advancing existing autoencoder methods.
The paper tackles the problem of generating samples along the data manifold rather than from parametric distributions like Gaussian, proposing a trajectory-aware principal manifold framework that extracts compact manifold representations and enables few-shot image generation, with results showing improved classification accuracy and smooth transformations.
Data augmentation for deep learning benefits model training, image transformation, medical imaging analysis and many other fields. Many existing methods generate new samples from a parametric distribution, like the Gaussian, with little attention to generate samples along the data manifold in either the input or feature space. In this paper, we verify that there are theoretical and practical advantages of using the principal manifold hidden in the feature space than the Gaussian distribution. We then propose a novel trajectory-aware principal manifold framework to restore the manifold backbone and generate samples along a specific trajectory. On top of the autoencoder architecture, we further introduce an intrinsic dimension regularization term to make the manifold more compact and enable few-shot image generation. Experimental results show that the novel framework is able to extract more compact manifold representation, improve classification accuracy and generate smooth transformation among few samples.