Back to projects page
Fast Poisson Blending using Multi-Splines
Richard Szeliski      Matthew Uyttendaele      Drew Steedly

Microsoft Research and Microsoft

International Conference on Computational Photography, April 2011
Top image is unblended collection of images, bottom is blended using the technique described in this paper.  This sequence is challenging as it contains exposure, and lighting changes as well as moving objects.  All of these are handled automatically to produce a seamlessly blended result.


We present a technique for fast Poisson blending and gradient domain compositing. Instead of using a single piecewise-smooth offset map to perform the blending, we associate a separate map with each input source image. Each individual offset map is itself smoothly varying and can therefore be represented using a low-dimensional spline. The resulting linear system is much smaller than either the original Poisson system or the quadtree spline approximation of a single (unified) offset map. We demonstrate the speed and memory improvements available with our system and apply it to large panoramas. We also show how robustly modeling the multiplicative gain rather than the offset between overlapping images leads to improved results, and how adding a small amount of Laplacian pyramid blending improves the results in areas of inconsistent texture.

Paper (ICCP 2011)

Click here(1.52 MB)
Supplemental Material
Click here

Back to projects page