摘要
医学图像融合是现代医学领域重要的研究方向之一。为了对病情做出准确全面的诊断及治疗,往往将来自不同医学图像的信息进行研究和分析,医学图像融合技术就是将多幅源图像信息融合到一幅图像中,以期获取更全面、清晰以及更丰富的图像信息。本文采用来自哈佛大学脑图像集中已配准好的CT和MRI、CT和PET两组对象,运用多种算法进行融合,重点研究基于小波变换的图像融合算法,将待融合的图像分解成高低频信息,然后通过特定的融合规则融合图像分解后的子信息,最后,通过逆变换得到最终的融合结果。本文通过评价标准证明了小波算法优于其他算法,此算法处理后的图像能得到良好的视觉效果和理想指标,在临床上具有重要的诊断参考价值。
Medical image fusion is the important research direction in the field of modern medicine. In order to make an accurate diagnosis and treatment to illness, many medical images＇ information should be analyzed. Medical image fusion technology is that mixes many source images＇ information into one image to obtain more comprehensive, clear and abundant image registration information. In the paper, with the two groups of CT and MRI, CT and PET of brain image by Harvard University, using a variety of fusion algorithm, focusing on researching the image fusion algorithm of wavelet transform. Firstly, we decompose the fused image into high frequency information, and then fuse the sub-information after decomposition through the specific fusion rules. Finally, obtain the final fused result by inverse transformation. By the subjective and objective evaluation, this algorithm is proved superior to the traditional algorithm, and it has the vital significance in clinical medicine.
出处
《世界复合医学》
2015年第3期277-282,共6页
World Journal of Complex Medicine
基金
国家重大科学仪器设备开发专项(2012YQ04014010)
黑龙江省自然科学基金(F201241)
关键词
小波变换
融合规则
医学图像
图像融合
wavelet transform
fusion rule
medical image
image fusion
