A Comprehensive Review of Multimodal Image Fusion Using Deep Learning Techniques

Authors

  • Samir B. Joshi Author
  • Amrutbhai N. Patel Author
  • Anand J. Patel Anand J. Patel Author

Keywords:

Multimodal Image Fusion, Deep learning, Architectures of DL, Applications of Multimodal image fusion

Abstract

Multimodal image fusion (MMIF), the process of combining complementary information from multiple imaging sources to enhance the quality and content of the resulting image, has gained significant attention across various fields including medical imaging, remote sensing, surveillance, robotics and automation, artificial intelligence, machine learning, and many more. Modalities and techniques of image fusion are of utmost importance. Deep learning (DL) architectures like Convolutional Neural Network, Generative Adversarial Networks, Auto-encoders, Transformers, etc. have revolutionized multimodal image fusion by providing powerful tools to learn complex mappings between heterogeneous data sources.

This paper presents a comprehensive review of recent advances in MMIF using DL methods. We discuss the fundamentals of MMIF, various DL architectures employed for fusion tasks, evaluation metrics, and various applications of MMIF. Additionally, we highlight challenges and future research scopes in this rapidly evolving area.

Downloads

Published

2025-01-01