arXiv 2602.030392026-02-05cs.CV

HP-GAN: 利用预训练网络提升GAN性能的FakeTwins与判别器一致性方法

HP-GAN: Harnessing pretrained networks for GAN improvement with FakeTwins and discriminator consistency

School of Electrical and Electronic Engineering, Yonsei University, Seoul, Republic of Korea（电气电子工程学院，延世大学，首尔，大韩民国）；Robotics Institute, Korea Institute of Science and Technology, Seoul, Republic of Korea（机器人研究所，韩国科学技术院，首尔，大韩民国）；Department of Oral and Maxillofacial Radiology, Yonsei University College of Dentistry, Seoul, Republic of Korea（口腔和颌面放射科，延世大学牙科学院，首尔，大韩民国）

Geonhui Son, Jeong Ryong Lee, Dosik Hwang

备注：: Accepted manuscript. This is the accepted version of the article published in Neural Networks
DOI：: 10.1016/j.neunet.2026.108666

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

HP-GAN通过FakeTwins和判别器一致性技术提升GAN性能，有效增强图像多样性和质量。

摘要：

生成对抗网络（GANs）在图像合成质量提升方面取得了显著进展。近期方法经常利用预训练网络计算感知损失或利用预训练特征空间。在本文中，我们通过引入创新的自监督学习技术和在GAN训练中强制判别器一致性来扩展预训练网络的能力。我们提出的方法称为HP-GAN，通过两种主要策略：FakeTwins和判别器一致性，有效利用神经网络先验。FakeTwins利用预训练网络作为编码器计算自监督损失，并通过生成图像训练生成器，从而生成更多多样性和高质量的图像。此外，我们引入了判别器之间的一致性机制，该机制评估从卷积神经网络（CNN）和视觉Transformer（ViT）特征网络提取的特征图。判别器一致性促进了判别器之间的协调学习，并通过对图像质量的评估进行对齐，从而增强训练的鲁棒性。我们在十七个数据集上进行了广泛评估，包括大量、少量和有限数据的场景，覆盖了各种图像领域，证明HP-GAN在Fréchet Inception距离（FID）方面始终优于当前最先进的方法，实现了图像多样性和质量的显著提升。代码可在：https://github.com/higun2/HP-GAN上获得。

英文摘要：

Generative Adversarial Networks (GANs) have made significant progress in enhancing the quality of image synthesis. Recent methods frequently leverage pretrained networks to calculate perceptual losses or utilize pretrained feature spaces. In this paper, we extend the capabilities of pretrained networks by incorporating innovative self-supervised learning techniques and enforcing consistency between discriminators during GAN training. Our proposed method, named HP-GAN, effectively exploits neural network priors through two primary strategies: FakeTwins and discriminator consistency. FakeTwins leverages pretrained networks as encoders to compute a self-supervised loss and applies this through the generated images to train the generator, thereby enabling the generation of more diverse and high quality images. Additionally, we introduce a consistency mechanism between discriminators that evaluate feature maps extracted from Convolutional Neural Network (CNN) and Vision Transformer (ViT) feature networks. Discriminator consistency promotes coherent learning among discriminators and enhances training robustness by aligning their assessments of image quality. Our extensive evaluation across seventeen datasets-including scenarios with large, small, and limited data, and covering a variety of image domains-demonstrates that HP-GAN consistently outperforms current state-of-the-art methods in terms of Fréchet Inception Distance (FID), achieving significant improvements in image diversity and quality. Code is available at: https://github.com/higun2/HP-GAN.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。