牙病图像特征提取算法的研究与设计

Research and Design on Medical Image Traits Draw-out Algorithm

为了解决现在许多自动分割方法不能满足医学图像处理实践要求的缺陷,考虑医学图像处理的特殊性,提出一种基于模糊连接的由用户参与控制、引导的交互式阈值分割算法.借助图像分割技术进行医学图像处理.采用牙病图像进行试验,结果表明此分割算法可以较好地描述图像中无法精确定义的区域,解决了医学图像处理实践中对分割准确性的要求,而且进一步解决了口腔图像中目标物体的边界不清晰而难以准确分割的问题.仿真实验结果证明这种算法的可行性和有效性.

To resolve the problem that automatic segment methods dissatisfied the demand of medical image, based on the particularity of medical image processing, an alternating threshold segment algorithm based on fuzzy connection by user＇s controlling and inducting was proposed. An experiment result of segment on tooth disease was given to show the segment algorithm could describe the undermined area in image, resolve the demand of veracity in medical image segment and the difficulty in exact segment on blurry edge of the focus in the nonnasality image. An emulational experiment was given to approve validity and feasibility of the algorithm.