Multiclass non-Adversarial Image Synthesis, with Application to Classification from Very Small Sample

TL;DR

This paper introduces COLA, a non-adversarial image synthesis method that outperforms GANs in small-data scenarios and enhances classification accuracy on challenging datasets.

Contribution

The paper presents COLA, a novel non-adversarial generative approach that overcomes GAN limitations and excels in small-sample image synthesis and classification tasks.

Findings

COLA generates diverse multi-class images without supervision.

COLA outperforms GANs in small-data regimes.

Augmenting datasets with COLA improves classification accuracy.

Abstract

The generation of synthetic images is currently being dominated by Generative Adversarial Networks (GANs). Despite their outstanding success in generating realistic looking images, they still suffer from major drawbacks, including an unstable and highly sensitive training procedure, mode-collapse and mode-mixture, and dependency on large training sets. In this work we present a novel non-adversarial generative method - Clustered Optimization of LAtent space (COLA), which overcomes some of the limitations of GANs, and outperforms GANs when training data is scarce. In the full data regime, our method is capable of generating diverse multi-class images with no supervision, surpassing previous non-adversarial methods in terms of image quality and diversity. In the small-data regime, where only a small sample of labeled images is available for training with no access to additional unlabeled data, our results surpass state-of-the-art GAN models trained on the same amount of data. Finally, when utilizing our model to augment small datasets, we surpass the state-of-the-art performance in small-sample classification tasks on challenging datasets, including CIFAR-10, CIFAR-100, STL-10 and Tiny-ImageNet. A theoretical analysis supporting the essence of the method is presented.