Automatic Correction of Internal Units in Generative Neural Networks

TL;DR

This paper introduces an automated method to identify and correct internal units in GAN generators responsible for artifacts, significantly improving image quality while maintaining original content.

Contribution

It proposes a novel approach to detect and sequentially correct artifact-causing units within GANs, enhancing image synthesis quality.

Findings

Outperforms baseline in FID-score

Achieves better human evaluation results

Effectively reduces visual artifacts

Abstract

Generative Adversarial Networks (GANs) have shown satisfactory performance in synthetic image generation by devising complex network structure and adversarial training scheme. Even though GANs are able to synthesize realistic images, there exists a number of generated images with defective visual patterns which are known as artifacts. While most of the recent work tries to fix artifact generations by perturbing latent code, few investigate internal units of a generator to fix them. In this work, we devise a method that automatically identifies the internal units generating various types of artifact images. We further propose the sequential correction algorithm which adjusts the generation flow by modifying the detected artifact units to improve the quality of generation while preserving the original outline. Our method outperforms the baseline method in terms of FID-score and shows satisfactory results with human evaluation.