Open Access

ARTICLE

Side-Scan Sonar Image Synthesis Based on CycleGAN with 3D Models and Shadow Integration

Byeongjun Kim^1,#, Seung-Hun Lee^2,#, Won-Du Chang^1,*

1 Department of Artificial Intelligence Convergence, Pukyong National University, Busan, 48513, Republic of Korea
2 Marine Domain Research Division, Korea Institute of Ocean Science and Technology (KIOST), Busan, 49111, Republic of Korea

* Corresponding Author: Won-Du Chang. Email:
# These authors contributed equally to this work

(This article belongs to the Special Issue: Machine Learning and Deep Learning-Based Pattern Recognition)

Computer Modeling in Engineering & Sciences 2025, 145(2), 1237-1252. https://doi.org/10.32604/cmes.2025.073530

Received 19 September 2025; Accepted 16 October 2025; Issue published 26 November 2025

Abstract

Side-scan sonar (SSS) is essential for acquiring high-resolution seafloor images over large areas, facilitating the identification of subsea objects. However, military security restrictions and the scarcity of subsea targets limit the availability of SSS data, posing challenges for Automatic Target Recognition (ATR) research. This paper presents an approach that uses Cycle-Consistent Generative Adversarial Networks (CycleGAN) to augment SSS images of key subsea objects, such as shipwrecks, aircraft, and drowning victims. The process begins by constructing 3D models to generate rendered images with realistic shadows from multiple angles. To enhance image quality, a shadow extractor and shadow region loss function are introduced to ensure consistent shadow representation. Additionally, a multi-resolution learning structure enables effective training, even with limited data availability. The experimental results show that the generated data improved object detection accuracy when they were used for training and demonstrated the ability to generate clear shadow and background regions with stability.

---

Keywords

---