基于在线字典学习的医学图像特征提取与融合.pdf

33卷3期
中国生物医学工程学报
Vol. 33 No. 3
2014年6月
Chinese Journal of Biomedical Engineering
基于在线字典学习的医学图像特征提取与融合
吴双邱天爽”高珊
(大连理工大学电子信息与电气工程学部,大连116024)
摘要:提出一种基于在线字典学习(ODL)的医学图像特征提取与融合的新算法。首先,采用大小为8像素x8
像素的滑动窗处理源图像,得到联合矩阵;通过OD.算法得到该联合矩阵的冗余字典,并利用最小角回归算法
(LARS)计算该联合矩阵的稀吭编码;将稀疏编码列向量的1范数作为稀疏编码的活动级测量准则,然后根据活动
级最大准则融合稀疏編;最后根据融合后的稀疏编码和冗余字典重构融合图像。实验图像为20位患者的已配
准脑部CT和MR图像,采用5种性能指标评价融合图像的质量,同两种流行的融合算法比较。结果显示,所提出
算法的各项客观指标均值最优, Piela指数、Q'指数、MI^"指数、BSSM指数和空间频率的均值分别为0.8004、
0.5524、3.6302、0.7269和31.9413,融合图像对比度、清晰度高,病灶的边缘清晰,运行速度较快,可以辅助医生
诊断和临床治疗。
关鍵词:图像融合;在线字典学习算法(ODL);最小角回归算法(IARS
中图分类号TP391;TN911.73文献标志码A文章编号0258-8021(2014)03-0283-06
Medical Image Features Extraction and Fusion Based on Online Dictionary Learning
WU Shuang QIU Tian-shuang GAO Shan
China Faculty of Electronic and Electrical Engineering, Dalian niversity of Technology, Dalian 116024, China)
Abstract: An image features extraction and fusion algorithm based on online dictionary learning(ODL) is
presented in this paper. Firstly, source images were combined into a joint matrix by the sliding window
technique, the size of the sliding window was x8, the over-complete dictionary was trained by ODI algorithm
and the sparse codes were acquired by LARS algorithm; the activity level measurement of sparse codes was the
Li norm of its vector, then, the sparse codes were fused by activity level maximum rule; finally, the fused
image was reconstructed by over-complete dictionary and fused sparse codes. Co-aligned medical images of
twenty patients were tested by experiments and the quality of the fused image was evaluated by five kinds of
commonly used objective criterions. Compared with the other two popular medical image fusion algorithms
objective criterions of the fusion result show the advantage of the proposed algorithm, the mean of Piella, Q
MI", BSSIM and space frequency index is 0. 4, 0. 5524, 3. 6302, 0.7269 and 31.941 3, the fusion
images of the proposed algorithm have high definition and contrast, clear texture and edge and fast speed
showing its application potentials of aiding clinical diagnoses and treatment
Key words: image fusion; online dictionary learning( ODL); least angle regression(LARS)algorithm
引言
性和冗余性的多模态医学图像生成更清晰、更完
整、更可靠医学图像。目前,医学图像融合主要可
医学图像融合技术是计算机处理、信息融合和分为3类,即像素级融合、特征级融合和决策级融
医学影像技术相结合的一种技术,它利用具有互补合。像素级融合是目前的研究重点,主要分为4类,
doi:10.3969/-isn.02588021.2014.03.04
收稿日期:2014-01-28,录用日期:201403-14
基金项月:国家自然科学基金(81241059,6172108);国家科技支撑计划项目(2012BAJ18BO6)
中国生物医学工程学会会员( Member of Chinese Society of Biomedical Engineering)
率通信作者(Correspondingauthor),E-mail:qiushi@du.edu.cn