基于结构自适应归一化卷积的磁共振图像超分辨率重建

目的 探讨一种针对磁共振图像超分辨率重建的有效算法.方法 根据图像间存在的微小结构差异,应用结构自适应归一化卷积算法,对重复扫描获取的磁共振图像进行超分辨率重建,同时运用其他4种常用超分辨率重建算法进行相同处理,计算峰值信噪比,比较重建效果.结果 结构自适应归一化卷积算法与其他算法相比,能够更好地保留磁共振图像的边缘和...

Super-resolution reconstruction of MR image based on structure-adaptive normalized convolution

Objective To explore an efficient super-resolution reconstruction algorithm for magnetic resonance image (MRI). Methods According to the tiny structural differences existing among MRI, the structure adaptive normalized convolution algorithm was applied to the super-resolution reconstruction of repeatedly scanned MRI. Other four kinds of algorithms were also used to compare the results of reconstruction. The peak signal to noise ratio (PSNR) between MRI and high-resolution images were calculated. Results Compared with other algorithms, structure-adaptive normalized convolution algorithm could better reserve the edge and the details of the images. Conclusion The algorithm, taking into consideration of local structure information to the reconstruction of high-resolution MRI, can significantly improve the quality of image.