计算机工程与应用 ›› 2015, Vol. 51 ›› Issue (19): 189-193.

• 图形图像处理 • 上一篇    下一篇

分布相关的曲波阈值超声图像去噪方法

徐  杰1,2,3,崔崤峣1,向永嘉1,吕铁军1,顾天明1   

  1. 1.中国科学院 苏州生物医学工程技术研究所,江苏 苏州 215163
    2.中国科学院大学,北京 100049
    3.中国科学院 长春光学精密机械与物理研究所,长春 130033
  • 出版日期:2015-09-30 发布日期:2015-10-13

Curvelet denoising method of ultrasound images based on probability distribution

XU Jie1,2,3, CUI Yaoyao1, XIANG Yongjia1, LV Tiejun1, GU Tianming1   

  1. 1.Suzhou Institute of Biomedical Engineering and Technology, CAS, Suzhou, Jiangsu 215163, China
    2.University of Chinese Academy of Sciences, Beijing 100049, China
    3.Changchun Institute of Optics, Fine Mechanics and Physics, CAS, Changchun 130033, China
  • Online:2015-09-30 Published:2015-10-13

摘要: 散斑噪声作为超声图像的主要噪声严重影响超声成像质量,滤除散斑噪声是超声图像处理过程中重要步骤。以曲波阈值去噪方法为基础,针对常用阈值函数中对较小曲波系数处理粗糙、不连续、收敛慢的缺点,通过分析实际超声图像中散斑噪声的分布,提出了具有与实际噪声分布相关特点的曲波阈值去噪方法。对比测试实验结果表明,曲波去噪方法相比其他去噪方法在不同噪声水平下均具有更加稳定优异的去噪性能,峰值性噪比提高了1~2 dB,平均结构相识度相比也有较大的提高。

关键词: 超声图像, 曲波, 散斑噪声去噪, 噪声分布

Abstract: Speckle noise is the main noise which seriously affects the quality of ultrasound images. Speckle noise filtering is an important step for medical ultrasound image processing. A new curvelet denoising method based on the probability distribution of speckle is proposed, to improve the performance of the threshold function’s continuity and astringency. Comparative results show that the proposed method has stable and outstanding performance at different noise levels. The method always has better Mean Structural Similarity Index Map(MSSIM) and increases the Peak Signal to Noise Ratio(PSNR) about 1~2 dB.

Key words: ultrasound image, curvelet, speckle noise reduction, noise probability distribution