计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (21): 195-198.

• 图形图像处理 • 上一篇    下一篇

全变差最小化在CT图像环状伪影去除中的应用

林志宏,李  光,罗守华   

  1. 东南大学 生物科学与医学工程学院,南京 210000
  • 出版日期:2015-11-01 发布日期:2015-11-16

Ring artifact correction via total variation minimization

LIN Zhihong, LI Guang, LUO Shouhua   

  1. School of Bio-medical Engineering, Southeast University, Nanjing 210000, China
  • Online:2015-11-01 Published:2015-11-16

摘要: 环状伪影是多种CT图像中普遍存在的一种伪影,主要由像元响应不理想等因素引起。全变差最小化是一种广泛使用的去噪方法,将全变差最小化应用于CT图像,在模拟投影数据中人为添加噪声并重建,使用全变差最小化进行处理,观察图像环状伪影的去除情况,并得出了图像峰值信噪比随全变差最小化次数的变化曲线。

关键词: 环状伪影, 全变差最小化, CT图像, 图像降噪

Abstract: Ring artifact is one of common artifacts appearing in CT images, which is mainly caused by the inconsistent response of detector pixels and other factors. Total variation minimization is a widely used image denoising method. In this paper, the total variation is applied on CT image to reduce ring artifact. Simulative projection data is added with noise. After reconstruction, total variation minimization is used to process the reconstruction image. The experiment demonstrates that total variation minimization can efficiently reduce ring artifact, and a curve of PSNR shows that the result can be controlled by total variation minimization numbers.

Key words: ring artifact, total variation minimization, CT image, image denoising