Slide 42 of 42

So two-dimensional opacity functions avoid that constraint by assinging opacity based on both the data value and and the gradient magnitude. Setting such an opacity function by hand is especially challenging because of the extra degrees of freedom involved. Fortunately, the techniques presented in the paper trivially generalize to produce two-dimensional opacity functions. This is posible with a postion function which now has the two-dimensional domain, while the user has the same one-dimensional boundary emphasis function. So the user is controlling a more powerful with opacity function, with the same intuitive interface.

Had these two slides been used, the next slide would have been the one with the four feet renderings (#36)