基于生成对抗网络的OCT图像去噪方法

OCT Image Denoising Method Based on Generative Confrontation Network

光学相干断层扫描(OCT)是用于监视和诊断眼部疾病的强大技术。然而,斑点噪声对OCT图像质量产生有害影响,这阻碍OCT辅助诊断的发展。受到深度学习快速发展的启发,本文将图像去噪问题视为图像的生成问题,并提出一种基于条件生成对抗网络(cGAN)的方法,通过学习有噪图像到无噪图像的映射来实现去噪过程。实验结果表明,本文提出的方法在视觉效果和指标上优于其他模型。

Optical Coherence Tomography(OCT)is a powerful technique for monitoring and diagnosing eye diseases.However,speckle noise has a harmful effect on OCT image quality,which hinders the development of OCT-assisted diagnosis.In this paper,we treated OCT image denoising as an image generation problem,and proposed a method based on conditional Generative Adversarial Network(cGAN),which realizes noise reduction by learning the mapping of noisy images to clean images.Experimental results show that the proposed method is superior to other models in terms of visual effects and indicators.