B超医学图像去噪模型的快速算法研究

Investigation on Fast Algorithm for B Ultrasonic Medical Image Denoising Model

数字图像处理技术在人们的现实生活中具有广泛的应用。深入研究基于变分偏微分方程的B超医学图像的噪声去除问题,既能丰富医学图像的处理方法,促进人们对图像的深入理解,又能为基于超声医学图像的诊断治疗提供帮助。针对B超医学图像中斑点噪声的去除问题,根据Jin等对去除超声图像中的乘性噪声变分模型进行的研究,结合该变分模型解的框式制约和Split Bregman算法,提出了针对B超医学图像中斑点噪声变分模型的快速数值求解算法,并且分析了所提算法的收敛性。对测试图像进行了数值仿真实验,并将文中提出的新方法与现有的算法进行了比较。实验结果表明,采用文中算法是可行有效的,能够在去除B超医学图像中斑点噪声的同时极大地缩短运算的时间。

Digital image processing technology has a wide range of applications in real life. B ultrasound medical image denoising problem is deeply studied based on variation and partial differential equation, which can not only make the method of medical image processing be rich,promotion of the deep understanding of the ultrasound image,but also be helpful for the diagnosis and treatment based on ultrasound image. For speckle noise removal problem of B ultrasound medical image, according to the study on the variational model of multiplica- tive noise in ultrasound images by Jin et al ,combination of the frame type constaxints of the model solution and the Split Bregman algo- rithm, a fast numerical algorithm is proposed for the variational model of removing the speckle noise in B ultrasound medical image, and its convergence is analyzed. The numerical simulation is carded out on the test image, and the new method is compared with the existing algorithm. The experimental results show that the algorithm proposed is feasible and effective, and the speckle noise in the B ultrasound medical image is removed and the computation time is greatly reduced.