CAJ | 학술논문

在医学图像配准优化算法中通常采用Powell法,由于基本Powell法随迭代次数的增加,搜索方向容易趋于线性相关,为此提出了一种改进的Powell法,该算法随着迭代的增加,搜索方向的共轭程度逐渐增强可避免线性相关.将边缘检测与最大互信息相结合,提高原有算法的性能,较准确地完成图像配准任务.对提出的配准算法进行了Matlab仿真实验并对仿真结果做了分析.
재의학도상배준우화산법중통상채용Powell법,유우기본Powell법수질대차수적증가,수색방향용역추우선성상관,위차제출료일충개진적Powell법,해산법수착질대적증가,수색방향적공액정도축점증강가피면선성상관.장변연검측여최대호신식상결합,제고원유산법적성능,교준학지완성도상배준임무.대제출적배준산법진행료Matlab방진실험병대방진결과주료분석.
The Powell algorithm is usually using in the medical image registration. In the search direction, the basic Powell algorithm becomes linear through the increase in the number of iterations. It presents a modified Powell algorithm. With the increase of iteration of the algorithm, the search direction conjugate levels gradually increase to avoid the linear eorrelation. The edge detection combined with the maximum mutual information improves the performance of the original algorithm that can achieve the task. It processes the Matlab simulation and analyze the results for the registration algorithms.