While reading a paper about multi-view stereo, Sometimes appear out "Unstructured image pair". But i cannot understand clearly. Could you explain to me ? Thanks.
I got the answer of this question.
In Multi-view stereo, many papers said that unstructured image pairs.
It means that unrectified images. So multi view cameras have different pose matrix.
On the other hands, stereo usually solved structured image equal to rectified image.
At this point, stereo depth estimation could solving this problem cost matching using disparity k between (i,j) and (i,j+k).