Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A survey on GANs for computer vision: Recent research, analysis and taxonomy

Guillermo Iglesias, Edgar Talavera, Alberto Díaz-Álvarez

TL;DR

This survey comprehensively analyzes Generative Adversarial Networks (GANs) for computer vision, detailing fundamental concepts, training challenges, and evaluation metrics, and then organizing the literature into architecture- and loss-function–based variants. It presents a taxonomy, timeline, and a broad view of applications from image synthesis to drug discovery, highlighting how advancements address mode collapse, instability, and training dynamics, while discussing the rising prominence of diffusion models and transformers as rivals. The work emphasizes practical aspects such as dataset quality, evaluation pitfalls, and societal implications, offering guidance for selecting architectures and losses for specific tasks. Overall, the paper maps the evolution of GANs over the past decade, identifies remaining gaps, and points to promising directions that balance fidelity, diversity, and training efficiency in real-world settings.

Abstract

In the last few years, there have been several revolutions in the field of deep learning, mainly headlined by the large impact of Generative Adversarial Networks (GANs). GANs not only provide an unique architecture when defining their models, but also generate incredible results which have had a direct impact on society. Due to the significant improvements and new areas of research that GANs have brought, the community is constantly coming up with new researches that make it almost impossible to keep up with the times. Our survey aims to provide a general overview of GANs, showing the latest architectures, optimizations of the loss functions, validation metrics and application areas of the most widely recognized variants. The efficiency of the different variants of the model architecture will be evaluated, as well as showing the best application area; as a vital part of the process, the different metrics for evaluating the performance of GANs and the frequently used loss functions will be analyzed. The final objective of this survey is to provide a summary of the evolution and performance of the GANs which are having better results to guide future researchers in the field.

A survey on GANs for computer vision: Recent research, analysis and taxonomy

TL;DR

This survey comprehensively analyzes Generative Adversarial Networks (GANs) for computer vision, detailing fundamental concepts, training challenges, and evaluation metrics, and then organizing the literature into architecture- and loss-function–based variants. It presents a taxonomy, timeline, and a broad view of applications from image synthesis to drug discovery, highlighting how advancements address mode collapse, instability, and training dynamics, while discussing the rising prominence of diffusion models and transformers as rivals. The work emphasizes practical aspects such as dataset quality, evaluation pitfalls, and societal implications, offering guidance for selecting architectures and losses for specific tasks. Overall, the paper maps the evolution of GANs over the past decade, identifies remaining gaps, and points to promising directions that balance fidelity, diversity, and training efficiency in real-world settings.

Abstract

In the last few years, there have been several revolutions in the field of deep learning, mainly headlined by the large impact of Generative Adversarial Networks (GANs). GANs not only provide an unique architecture when defining their models, but also generate incredible results which have had a direct impact on society. Due to the significant improvements and new areas of research that GANs have brought, the community is constantly coming up with new researches that make it almost impossible to keep up with the times. Our survey aims to provide a general overview of GANs, showing the latest architectures, optimizations of the loss functions, validation metrics and application areas of the most widely recognized variants. The efficiency of the different variants of the model architecture will be evaluated, as well as showing the best application area; as a vital part of the process, the different metrics for evaluating the performance of GANs and the frequently used loss functions will be analyzed. The final objective of this survey is to provide a summary of the evolution and performance of the GANs which are having better results to guide future researchers in the field.
Paper Structure (36 sections, 27 equations, 11 figures, 3 tables)

This paper contains 36 sections, 27 equations, 11 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Structure of this survey
  4. gan
  5. Definition and structure
  6. Common problems
  7. Mode collapse
  8. Gradient vanishing
  9. Instability
  10. Stopping problem
  11. Evaluation metrics
  12. and its variants
  13. Perceptual path length
  14. variants
  15. Architecture optimization
  16. ...and 21 more sections

Figures (11)

  • Figure 1: Architecture of a model.
  • Figure 2: Structure of the proposed architecture of the GANILLA. Figure based on Reference hicsonmez2020ganilla.
  • Figure 3: Training schedule of ProGAN. Figure based on Reference karras2018progressive.
  • Figure 4: Training methodology of . Figure based on Reference liu2021dynamically.
  • Figure 5: Self attention layer of . Figure based on Reference zhang2019self.
  • ...and 6 more figures