question and example of point density based on image values

I have been working on trying to fit points to an image based upon image sampler values, where darker areas yield more points and lighter areas less points. There is a previous discussion here http://www.grasshopper3d.com/forum/topics/image-sampler-point-density

The following link is the solution I am using for the images seen above which somewhat works. 

stippling_works.gh

From what I've read they usually do this sort of thing by using a weighted voronoi and then filling in the cells with points based upon the resulting area. I tried this but the way I was processing the weighted voronoi can get very slow. What I would like to try but I haven't been able to figure it out is to measure a random pointset against and image and based upon those values pull all other points proportional to the values made from the image sampler. This would end up being an attractor type function capable of using values to control the attraction force. This seems like it could be an interesting way to control point density.