Dictionaries are known tools used in different branches of image processing like edge detection, inpainting and, etc. Segmentation is the task of extracting an object as the part of a particular image. The common drawback of different segmentation methods is that they perform the extraction task incompletely. Tasks like edge detection, denoising and smoothing, as the parts of segmentation, can be done through applying the dictionaries. In this paper, we propose three new contrast stretching function. Based on one of the stretching functions and shearlets as a dictionary, we improved the previous version of a method that has been used in binary segmentation for magnetic resonance angiography images (MRI). We also introduce a three-stage binary image segmentation algorithm for vessel segmentation in MRI images. There are some disadvantages in recent proposed methods when dealing with extracting vessels of medical images. Our algorithm does the task with a more accurate extraction in detecting vessels having low intensity and weak edges in MRI.

词典是在图像处理的不同分支（如边缘检测，修复等）中使用的已知工具。分割是提取对象作为特定图像的一部分的任务。不同分割方法的共同缺点是它们无法完全执行提取任务。边缘检测，去噪和平滑等任务是分割的一部分，可以通过应用字典来完成。在本文中，我们提出了三种新的对比度拉伸功能。基于伸展函数和小波片之一作为字典，我们改进了先前用于磁共振血管造影图像（MRI）的二值分割方法的版本。我们还介绍了用于MRI图像血管分割的三阶段二值图像分割算法。在处理医学图像的提取血管时，最近提出的方法存在一些缺点。我们的算法以更精确的提取来完成任务，以检测MRI中强度低且边缘薄弱的血管。