基于插值与剪切波融合的图像超分辨率重建

Image Super-resolution Reconstruction Based on Interpolation and Shearlet Fusion

针对单幅图像超分辨率重建问题,提出一种基于软判决自适应(SAI)-双三次(Bicubic)插值与平移不变剪切波融合的超分辨率重建算法。对源图像分别进行SAI插值和Bicubic插值,采用平移不变剪切波变换对2幅插值图像进行多尺度、多方向分解,得到低频及高频子带,对于低频子带,根据区域系数方差确定模糊相似度,结合改进的S函数确定自适应加权融合规则,对于高频子带,采用新改进拉普拉斯能量和与加权平均相结合的融合规则进行处理,将得到的融合系数进行剪切波逆变换,从而得到高分辨率重建图像。实验结果表明,与原有的SAI插值算法相比,该算法能提升重建图像的清晰度及峰值信噪比。

For a single image Super-resolution（ SR） reconstruction problem,a novel image SR algorithm based on Soft-decision Adaptive Interpolation （ SAI ）-Bicubic interpolation and Shift-invariant Shearlet Transform （ SIST ） fusion is proposed. For each source image is separately interpolated by SAI and Bicubic interpolation,and the SIST is adopted to decompose the two interpolated images in different scales and directions,and the low-frequency and high-frequency sub-band coefficients of the two images are obtained. For the low frequency sub-band coefficients,according to the regional variance to determine the fuzzy similarity,a adaptive weighted fusion rule combined with improved sigmoid function is presented. For the high frequency sub-band coefficients,it uses a new Sum-modified Laplacian（ SML） and is combined with the weighted average fusion rule. The high resolution image is obtained by performing the inverse SIST on the combined coefficients. Compared with the SAI,the imposed algorithm has very good effect on improving the clarity of the reconstructed image and Peak Signal to Noise Ratio（ PSNR） .