图像相似度计算MATLAB代码-TextureDeblur:用于3D重建的纹理去毛刺

Matlab code for the paper J. Wu, H. Chen, X. Liu, L. Cao, X. Peng, and G. Jin, "Unsupervised texture reconstruction method using bidirectional similarity function for 3-D measurements," Optics Communications 439, 85-93 (2019). It is used for eliminating artifacts like blurring, ghosting and color discontinuity in texture image of 3-D measurements. 1. Run the start, the algorithm will be unsupervised performed and generate optimized texture images saved in '.\image'. 2. Description of the functions targetTexture.m - texture image optimization function savePp.m - texture mappingng to obtain UV coordinates immapping.m - pixel remapping function, which implement image projection between different view. weight.m - weight calculation function nnmex.mexw64 / votemex.mexw64 - PatchMatch algorithm proposed by Barnes et al. It is a compiled machine code for a Windows 64-bit setup. see more details in nnmex.m and votemex.m Others are helper functions.