Hi,
is there any way to use masks for texturing?
I explain my problem: I have two sets of images of the model. One set with lots of overlapped photographies but with bad illumination, and the other one with a few photographies with good illumination but bad overlapping. I’m using both groups for alignment and mesh generation and only the second one for texturing.
However, the overlapping on the second group is so bad, that there are some small parts missing. I know best solution is taking more pictures with good illumination to cover this missing parts, but this is not possible.
So, I thought to use some of the bad illuminated photographies to fill these parts. But, although they fill this missing parts in a relatively good way (after some postprocessing of the images), they make other parts worse: parts that were very well represented in the second group starts to have noise or even important changes in color.
I have tried visibility-based and photo-consistency algorithm and the results are very bad in both cases. So, I can only imagine some texture masking to solve the situation. But, as far as I know, masks in RC only affect to alignment and meshing… Is that right? Any other idea?
Thanks!