cGANs with Conditional Convolution Layer

TL;DR

This paper introduces a novel conditional convolution layer for cGANs that adjusts weights based on conditions, enabling more effective learning of condition-specific features and improving image generation quality.

Contribution

The paper proposes a new conditional convolution layer that dynamically adjusts weights based on conditions, enhancing cGANs' ability to generate condition-specific images with a single generator.

Findings

Higher quality conditional image generation on CIFAR, LSUN, and ImageNet datasets.

Effective handling of condition-specific features with a single generator.

Improved performance over standard convolution in cGANs.

Abstract

Conditional generative adversarial networks (cGANs) have been widely researched to generate class conditional images using a single generator. However, in the conventional cGANs techniques, it is still challenging for the generator to learn condition-specific features, since a standard convolutional layer with the same weights is used regardless of the condition. In this paper, we propose a novel convolution layer, called the conditional convolution layer, which directly generates different feature maps by employing the weights which are adjusted depending on the conditions. More specifically, in each conditional convolution layer, the weights are conditioned in a simple but effective way through filter-wise scaling and channel-wise shifting operations. In contrast to the conventional methods, the proposed method with a single generator can effectively handle condition-specific characteristics. The experimental results on CIFAR, LSUN and ImageNet datasets show that the generator with the proposed conditional convolution layer achieves a higher quality of conditional image generation than that with the standard convolution layer.