Generative Adversarial Networks using Adaptive Convolution

TL;DR

This paper introduces an adaptive convolution method for GANs that learns local upsampling algorithms, significantly improving image generation quality and achieving state-of-the-art results on CIFAR-10 and STL-10 datasets.

Contribution

The paper presents a novel adaptive convolution technique for GANs that replaces fixed upsampling methods, enhancing model flexibility and performance.

Findings

Improved image quality on CIFAR-10 dataset.

Achieved state-of-the-art results on CIFAR-10 and STL-10.

Significant performance margin over baseline models.

Abstract

Most existing GANs architectures that generate images use transposed convolution or resize-convolution as their upsampling algorithm from lower to higher resolution feature maps in the generator. We argue that this kind of fixed operation is problematic for GANs to model objects that have very different visual appearances. We propose a novel adaptive convolution method that learns the upsampling algorithm based on the local context at each location to address this problem. We modify a baseline GANs architecture by replacing normal convolutions with adaptive convolutions in the generator. Experiments on CIFAR-10 dataset show that our modified models improve the baseline model by a large margin. Furthermore, our models achieve state-of-the-art performance on CIFAR-10 and STL-10 datasets in the unsupervised setting.