ICSO: A Novel Hybrid Evolutionary Approach with Crisscross and Perturbation Mechanisms for Optimizing Generative Adversarial Network Latent Space

Hybrid evolutionary approaches have gained significant attention for solving complex optimization problems, but their potential for optimizing the low-dimensional latent space of generative adversarial networks (GANs) remains underexplored. This paper proposes a novel improved crisscross optimizatio...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhihui Chen, Ting Lan, Zhanchuan Cai, Zonglin Liu, Renzhang Chen
Format: Article
Language:English
Published: MDPI AG 2025-05-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/10/5228
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850257927084965888
author Zhihui Chen
Ting Lan
Zhanchuan Cai
Zonglin Liu
Renzhang Chen
author_facet Zhihui Chen
Ting Lan
Zhanchuan Cai
Zonglin Liu
Renzhang Chen
author_sort Zhihui Chen
collection DOAJ
description Hybrid evolutionary approaches have gained significant attention for solving complex optimization problems, but their potential for optimizing the low-dimensional latent space of generative adversarial networks (GANs) remains underexplored. This paper proposes a novel improved crisscross optimization (ICSO) algorithm, a hybrid evolutionary approach that integrates crisscross optimization and perturbation mechanisms to find the suitable latent vector. The ICSO algorithm treats the quality and diversity as separate objectives, balancing them through a normalization strategy, while a gradient regularization term (i.e., GP) is introduced into the discriminator’s objective function to stabilize training and mitigate gradient-related issues. By combining the global and local search capabilities of particle swarm optimization (PSO) with the rapid convergence of crisscross optimization, ICSO efficiently explores and exploits the latent space. The extensive experiments demonstrate that ICSO outperforms state-of-the-art algorithms in optimizing the latent space of various classical GANs across multiple datasets. Furthermore, the practical applicability of ICSO is validated through its integration with StyleGAN3 for generating unmanned aerial vehicle (UAV) images, showcasing its effectiveness in real-world engineering applications. This work not only advances the field of GAN optimization but also provides a robust framework for applying hybrid evolutionary algorithms to complex generative modeling tasks.
format Article
id doaj-art-e4e8bef3046b435c9f83a79cde9e49f0
institution OA Journals
issn 2076-3417
language English
publishDate 2025-05-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-e4e8bef3046b435c9f83a79cde9e49f02025-08-20T01:56:17ZengMDPI AGApplied Sciences2076-34172025-05-011510522810.3390/app15105228ICSO: A Novel Hybrid Evolutionary Approach with Crisscross and Perturbation Mechanisms for Optimizing Generative Adversarial Network Latent SpaceZhihui Chen0Ting Lan1Zhanchuan Cai2Zonglin Liu3Renzhang Chen4School of Computer Science and Engineering, Macau University of Science and Technology, Macau 999078, ChinaSchool of Computer Science and Engineering, Macau University of Science and Technology, Macau 999078, ChinaSchool of Computer Science and Engineering, Macau University of Science and Technology, Macau 999078, ChinaLANGO Tech Co., Ltd., Guangzhou 510000, ChinaThe Modern Educational Technology Center, Jinan University, Zhuhai 519070, ChinaHybrid evolutionary approaches have gained significant attention for solving complex optimization problems, but their potential for optimizing the low-dimensional latent space of generative adversarial networks (GANs) remains underexplored. This paper proposes a novel improved crisscross optimization (ICSO) algorithm, a hybrid evolutionary approach that integrates crisscross optimization and perturbation mechanisms to find the suitable latent vector. The ICSO algorithm treats the quality and diversity as separate objectives, balancing them through a normalization strategy, while a gradient regularization term (i.e., GP) is introduced into the discriminator’s objective function to stabilize training and mitigate gradient-related issues. By combining the global and local search capabilities of particle swarm optimization (PSO) with the rapid convergence of crisscross optimization, ICSO efficiently explores and exploits the latent space. The extensive experiments demonstrate that ICSO outperforms state-of-the-art algorithms in optimizing the latent space of various classical GANs across multiple datasets. Furthermore, the practical applicability of ICSO is validated through its integration with StyleGAN3 for generating unmanned aerial vehicle (UAV) images, showcasing its effectiveness in real-world engineering applications. This work not only advances the field of GAN optimization but also provides a robust framework for applying hybrid evolutionary algorithms to complex generative modeling tasks.https://www.mdpi.com/2076-3417/15/10/5228crisscross optimizationevolutionary optimizationgenerative adversarial networklatent spaceparticle swarm optimization
spellingShingle Zhihui Chen
Ting Lan
Zhanchuan Cai
Zonglin Liu
Renzhang Chen
ICSO: A Novel Hybrid Evolutionary Approach with Crisscross and Perturbation Mechanisms for Optimizing Generative Adversarial Network Latent Space
Applied Sciences
crisscross optimization
evolutionary optimization
generative adversarial network
latent space
particle swarm optimization
title ICSO: A Novel Hybrid Evolutionary Approach with Crisscross and Perturbation Mechanisms for Optimizing Generative Adversarial Network Latent Space
title_full ICSO: A Novel Hybrid Evolutionary Approach with Crisscross and Perturbation Mechanisms for Optimizing Generative Adversarial Network Latent Space
title_fullStr ICSO: A Novel Hybrid Evolutionary Approach with Crisscross and Perturbation Mechanisms for Optimizing Generative Adversarial Network Latent Space
title_full_unstemmed ICSO: A Novel Hybrid Evolutionary Approach with Crisscross and Perturbation Mechanisms for Optimizing Generative Adversarial Network Latent Space
title_short ICSO: A Novel Hybrid Evolutionary Approach with Crisscross and Perturbation Mechanisms for Optimizing Generative Adversarial Network Latent Space
title_sort icso a novel hybrid evolutionary approach with crisscross and perturbation mechanisms for optimizing generative adversarial network latent space
topic crisscross optimization
evolutionary optimization
generative adversarial network
latent space
particle swarm optimization
url https://www.mdpi.com/2076-3417/15/10/5228
work_keys_str_mv AT zhihuichen icsoanovelhybridevolutionaryapproachwithcrisscrossandperturbationmechanismsforoptimizinggenerativeadversarialnetworklatentspace
AT tinglan icsoanovelhybridevolutionaryapproachwithcrisscrossandperturbationmechanismsforoptimizinggenerativeadversarialnetworklatentspace
AT zhanchuancai icsoanovelhybridevolutionaryapproachwithcrisscrossandperturbationmechanismsforoptimizinggenerativeadversarialnetworklatentspace
AT zonglinliu icsoanovelhybridevolutionaryapproachwithcrisscrossandperturbationmechanismsforoptimizinggenerativeadversarialnetworklatentspace
AT renzhangchen icsoanovelhybridevolutionaryapproachwithcrisscrossandperturbationmechanismsforoptimizinggenerativeadversarialnetworklatentspace