Part 1.4: Multiresolution Blending

Image set 1 - Oraple

I used a sigma of 50 for the Gaussian filter on the mask, and a sigma of 4 for the filter on each image.
not found not found not found

Image set 2 - Sportball

I used a sigma of 40 for the Gaussian filter on the mask, and a sigma of 2 for the filter on each image. It did not turn out as well as the other examples. I think this is because the original images are too detailed and sharp to blur convincingly.
not found not found not found

Image set 3 - Space

I used a sigma of 20 for the Gaussian filter on the mask, and a sigma on 3 for the filter on each image. It turned out fairly well, but I think the shading behind Jupiter is not entirely convincing. I revisited this example in Part 2.2, with better results.
not found not found not found

Laplacian stacks for Oraple

By looking at the Laplacian stacks of the Oraple, and the masked input images, we can see how the Laplacians of the oraple were formed as linear combinations of the Laplacians of the masked images.
not found
not found not found not found not found not found
Masked apple
not found
not found not found not found not found not found
Masked orange
not found
not found not found not found not found not found
Blended image

Bells and Whistles - Adding Color

By blending each RGB channel separately, we can blend color images as easily as we did black-and-white images. The color images turned out more convincingly than the black-and-white images, in my opinion.
not found not found not found