Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GAN-GA: A Generative Model based on Genetic Algorithm for Medical Image Generation

M. AbdulRazek, G. Khoriba, M. Belal

TL;DR

Medical image synthesis is hindered by small, imbalanced datasets and privacy constraints. The authors propose InfoGAN-GA, a generative model that embeds a Genetic Algorithm within InfoGAN to guide generation, accelerate learning, and improve fidelity and diversity of synthetic medical images. Evaluated on a small Acute Lymphoblastic Leukemia dataset, InfoGAN-GA achieves about a 6.8% improvement in Frechet Inception Distance over InfoGAN, with stronger gains in early training epochs, indicating faster and more stable convergence. These results suggest evolutionary optimization can enhance medical image synthesis and data augmentation, with potential applicability to other imaging domains and broader benchmarking against state-of-the-art methods.

Abstract

Medical imaging is an essential tool for diagnosing and treating diseases. However, lacking medical images can lead to inaccurate diagnoses and ineffective treatments. Generative models offer a promising solution for addressing medical image shortage problems due to their ability to generate new data from existing datasets and detect anomalies in this data. Data augmentation with position augmentation methods like scaling, cropping, flipping, padding, rotation, and translation could lead to more overfitting in domains with little data, such as medical image data. This paper proposes the GAN-GA, a generative model optimized by embedding a genetic algorithm. The proposed model enhances image fidelity and diversity while preserving distinctive features. The proposed medical image synthesis approach improves the quality and fidelity of medical images, an essential aspect of image interpretation. To evaluate synthesized images: Frechet Inception Distance (FID) is used. The proposed GAN-GA model is tested by generating Acute lymphoblastic leukemia (ALL) medical images, an image dataset, and is the first time to be used in generative models. Our results were compared to those of InfoGAN as a baseline model. The experimental results show that the proposed optimized GAN-GA enhances FID scores by about 6.8\%, especially in earlier training epochs. The source code and dataset will be available at: https://github.com/Mustafa-AbdulRazek/InfoGAN-GA.

GAN-GA: A Generative Model based on Genetic Algorithm for Medical Image Generation

TL;DR

Medical image synthesis is hindered by small, imbalanced datasets and privacy constraints. The authors propose InfoGAN-GA, a generative model that embeds a Genetic Algorithm within InfoGAN to guide generation, accelerate learning, and improve fidelity and diversity of synthetic medical images. Evaluated on a small Acute Lymphoblastic Leukemia dataset, InfoGAN-GA achieves about a 6.8% improvement in Frechet Inception Distance over InfoGAN, with stronger gains in early training epochs, indicating faster and more stable convergence. These results suggest evolutionary optimization can enhance medical image synthesis and data augmentation, with potential applicability to other imaging domains and broader benchmarking against state-of-the-art methods.

Abstract

Medical imaging is an essential tool for diagnosing and treating diseases. However, lacking medical images can lead to inaccurate diagnoses and ineffective treatments. Generative models offer a promising solution for addressing medical image shortage problems due to their ability to generate new data from existing datasets and detect anomalies in this data. Data augmentation with position augmentation methods like scaling, cropping, flipping, padding, rotation, and translation could lead to more overfitting in domains with little data, such as medical image data. This paper proposes the GAN-GA, a generative model optimized by embedding a genetic algorithm. The proposed model enhances image fidelity and diversity while preserving distinctive features. The proposed medical image synthesis approach improves the quality and fidelity of medical images, an essential aspect of image interpretation. To evaluate synthesized images: Frechet Inception Distance (FID) is used. The proposed GAN-GA model is tested by generating Acute lymphoblastic leukemia (ALL) medical images, an image dataset, and is the first time to be used in generative models. Our results were compared to those of InfoGAN as a baseline model. The experimental results show that the proposed optimized GAN-GA enhances FID scores by about 6.8\%, especially in earlier training epochs. The source code and dataset will be available at: https://github.com/Mustafa-AbdulRazek/InfoGAN-GA.
Paper Structure (11 sections, 3 equations, 10 figures, 2 tables, 1 algorithm)

This paper contains 11 sections, 3 equations, 10 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Proposed InfoGAN-GA: Generative Model embedded with Genetic Algorithm
  4. Generative Adversarial Networks (GANs)
  5. Information Maximizing Generative Adversarial Nets (InfoGAN)
  6. Genetic Algorithm
  7. InfoGAN-GA: The Proposed Approach
  8. Results and Discussions
  9. Experimental Settings
  10. Evaluation Results
  11. Conclusion

Figures (10)

  • Figure 1: Original Generative Adversarial Nets (GAN)
  • Figure 2: Generative Model embedded with Genetic Algorithm (InfoGAN-GA) Architecture. Latent vector z as input to Generator. $GA[G(Z, c)]$ is the output of genetic algorithm. Q is an auxiliary network attached to the second to last layer of the discriminator inspired from original InfoGAN.
  • Figure 3: Acute lymphoblastic leukemia (ALL) image dataset sample images 28x28.
  • Figure 4: Illustration of Original vs. Generated images from InfoGAN and InfoGAN-GA versions 28x28.
  • Figure 5: Comparison between synthesized images after first epoch for base InfoGAN, InfoGAN-GA-WoC, and InfoGAN-GA-WC models.
  • ...and 5 more figures