Optical-to-SAR image translation based on diffusion Schrödinger bridge

Junyu Wang; Hao Sun; Tao Tang; Lingjun Zhao; Lin Lei; Kefeng Ji

doi:10.1117/12.3037223

4 December 2024 Optical-to-SAR image translation based on diffusion Schrödinger bridge

Junyu Wang, Hao Sun, Tao Tang, Lingjun Zhao, Lin Lei, Kefeng Ji

Author Affiliations +

Proceedings Volume 13283, Conference on Spectral Technology and Applications (CSTA 2024); 132834D (2024) https://doi.org/10.1117/12.3037223
Event: Conference on Spectral Technology and Applications (CSTA 2024), 2024, Dalian, China

Abstract

In the field of synthetic aperture radar (SAR) target recognition, the recognition of aircraft targets has consistently presented a significant challenge. This is primarily due to the difficulty in obtaining SAR data of aircraft. In contrast, optical images of the target are easy to obtain. Consequently, the prevailing approach has been to translate target optical images into SAR images for target recognition. With the advancement of diffusion models, the quality and diversity of generated samples have continued to improve, rendering it possible to employ diffusion models for the translation of optical images to SAR. Additionally, the Schrödinger bridge constructs a transition between two arbitrary distributions, which can be utilized in the diffusion model to achieve optimal transport with two distributions. This paper proposes an innovative solution to achieve the translation from optical to SAR images by diffusion model based on the Schrödinger bridge. Concurrently, there is an unavoidable domain shift between the generated image and the measured SAR image. In order to narrow this domain shift, a domain adaptation method is employed. The experimental results demonstrate that the data generated by diffusion can effectively assist in the recognition of three types of aircraft targets, with an accuracy rate of 71.11%.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Junyu Wang, Hao Sun, Tao Tang, Lingjun Zhao, Lin Lei, and Kefeng Ji "Optical-to-SAR image translation based on diffusion Schrödinger bridge", Proc. SPIE 13283, Conference on Spectral Technology and Applications (CSTA 2024), 132834D (4 December 2024); https://doi.org/10.1117/12.3037223

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE