jzhangbs
jzhangbs
you can find "pair.txt" in the data folders
it is possible. the time for pair-wise regularization is linear wrt number of views.
depth filtering is the only post processing before fusion. some noise may not be visible because of the white background. you can also refer to the visualization on the benchmark...
short range. the camera files are provided by MVSNet
it is not used. you can generate a dummy mask by ones_like, with the same size as the depth map
您好，MVSDF预计在12月放出代码
Hi, thanks for providing this new work. I have quickly go through the paper. It seems that the structure map is used to get rid of RGB information that is...
您好，“使用了 MVS的DTU数据集进行了训练”具体是怎么操作的？应该是用官网下载的rectified image和我们提供的其他文件来训练
mvsnet repo提供的小图是crop过的，直接输入网络就可以，所以这边应该把resize和crop的size设置成不变（640x512）。NanError的原因就比较复杂，先试一下上面这个建议，再看看Nan的问题是否依然存在
可以先检查一下depth map是否正确，可以用提供的plot_pfm.py