基于在线字典学习和脉冲耦合神经网络的脑图像融合

医学图像融合是医学影像和放射医学等领域的研究热点之一，广受医学界和工程界重视。提出一种基于在线字典学习(ODL)和脉冲耦合神经网络(PCNN)的脑部CT和MR图像融合新算法。首先，利用滑动窗技术将源图像分块，使用ODL算法和最小角回归算法(LARS)得到各图像块对应列向量的稀疏编码；其次，将稀疏编码作为脉冲耦合神经网络的外部输入刺激信号进行迭代处理，根据点火次数确定融合系数；最后，根据融合系数和学习字典重构融合图像。基于哈佛医学院的10组脑部CT和MR数据，将所提出算法同基于KSVD的融合算法、基于ODL的融合算法、基于NSCT的融合算法比较。实验结果显示：综合考虑主观视觉效果和客观评价指标，该算法性能整体优于其他算法，客观参数指标BSSIM、MI、Piella、SF、STD、Q^AB/F的均值分别为0.751 2、3.769 6、0.697 1、29.526 7、90.090 6、0.570 7，可以提供丰富的信息来辅助医生分析病变体，提高临床医疗诊断的准确性和治疗规划的科学性。

Brain Image Fusion Based on Online Dictionary Learning and Pulse Coupled Neural Network

Medical image fusion is an important issue in the field of medical imaging and radiation medicine, and is widely recognized by medical and engineering fields. In this study, a new algorithm of brain CT and MR image fusion was proposed based on the online dictionary learning (ODL) and pulse coupled neural network (PCNN). First, the source images were divided into patches using the sliding window technology, and then the ODL algorithm and LARS algorithm were used to get the sparse coding coefficient of corresponding column vector of each image patch. Second, the sparse coding coefficient was selected as the external stimulus input of the pulse coupled neural network for iterative processing, and the fusion coefficient was determined by firing times. At last, the fusion image was reconstructed according to the fusion coefficients and the learned dictionary. Based on CT and MR brain data of 10 groups that from the Harvard Medical School, and  compared with the KSVD-based fusion method, the ODL-based fusion method and the NSCT-based fusion method, the experimental results showed that the proposed algorithm was superior to the other algorithms in terms of subjective visual effects and objective evaluation index. The average of objective parameters BSSIM, MI, Piella, SF, STD, Q^AB/F were 0.751 2, 3.769 6, 0.697 1, 29.526 7, 90.090 6, 0.570 7, and the fusion results can provide abundant information to help medical doctors to analyze the pathological tissue, thus it can improve the accuracy of clinical medical diagnosis and make efficient treatment plans.