The stereo pair we're gonna work on comes from Peter Simcoe, who was kind enough to let me experiment with his Fuji W3 pictures.
This is the left image.
That is the right image. This is a stereo pair from which a good depth map should be easily obtained.
Let's get a depth map using Depth Map Automatic Generator 7 (DMAG7).
This is the depth map generated by DMAG7 using the following parameters: min disparity = -32, max disparity = 6, spatial sample rate = 8, range spatial rate = 32, patch radius = 12, lambda = 0.01, max L-BFGS iterations = 500, hash table size = 10,000. You might be able to get a better depth map by fiddling around with the parameters but that's good enough for me as is. For the min and max disparities, I didn't even bother using Disparity Finder 2 (DF2), I just put some values until I got white for whatever is the most in front and black for whatever is the most in the back. Since that was taken with a Fuji W3 without parallax fussing around, the foreground is probably at around 0 disparity and the background at -width/30 = -27 (width = 800).
This is the 3d wiggle generated by Wiggle Maker.
Automatic depth map generation, stereo matching, multi-view stereo, Structure from Motion (SfM), photogrammetry, 2d to 3d conversion, etc. Check the "3D Software" tab for my free 3d software. Turn photos into paintings like impasto oil paintings, cel shaded cartoons, or watercolors. Check the "Painting Software" tab for my image-based painting software. Problems running my software? Send me your input data and I will do it for you.


This I really like, really delicious result
ReplyDelete