We're offering the Project Mosul initiative the help of the Sketchfab community's expertise - see details here.
As a first project for us, we're interested in your opinion of how we can improve the reconstruction results. The first 3D reconstruction, which was a 'proof of concept', was based on only 16 pictures from different sources. Made with different cameras under different lighting conditions, it's quite a challenge to get a good result from them.
And here's the reconstruction after masking, and using Photoshop auto-levels to make the colors more consistent:
My first question to you: how can we get a better result from such a set of photos? Which steps are necessary?