CAJ | 학술논문

对 ASIFT 算法的原理进行了分析,针对医学图像配准鲁棒性强、准确性高和低耗时率的要求,设计出基于ASIFT 的医学图像配准算法.该算法首先通过 ASIFT 算法提取图像特征点,接着用欧式距离筛选出匹配的特征点,最后实现参考图像与浮动图像之间的配准.该算法较好地解决了其他同类型算法中存在的提取的特征点数量少、特征点匹配的精确度不高、不能对扭曲变形的仿射图像配准等问题.实验结果表明,该算法不仅提高了配准的精确度和准确性,也提高了配准的稳定性和可靠性.
대 ASIFT 산법적원리진행료분석,침대의학도상배준로봉성강、준학성고화저모시솔적요구,설계출기우ASIFT 적의학도상배준산법.해산법수선통과 ASIFT 산법제취도상특정점,접착용구식거리사선출필배적특정점,최후실현삼고도상여부동도상지간적배준.해산법교호지해결료기타동류형산법중존재적제취적특정점수량소、특정점필배적정학도불고、불능대뉴곡변형적방사도상배준등문제.실험결과표명,해산법불부제고료배준적정학도화준학성,야제고료배준적은정성화가고성.
This paper analyzes the principle of ASIFT algorithm.For medical image registration of high robustness, high accu-racy and low consumption rate, this paper proposes a medical image registration algorithm based on ASIFT .This algorithm extracts feature points from images by the ASIFT algorithm, then matches the feature points by the Euclidean distance, and achieves the medical image registration between two images at last.The algorithm effectively deals with the problems such as less feature points collected, inaccuracy matching appeared in feature points, being incapable to registration on distorted affine image, etc.,which exist in other algorithms.The experimental result verifies that the algorithm not only improves the registration accuracy and preci-sion, but also enhances the stability and reliability of the registration.