Drawing Parrots with Charcoal

We present an algorithm to convert color images to gray scale that ensures the separation between iso-luminance color regions and improves the contrast in the resulting image. The algorithm calculates a local vector in the neighborhood of a pixel with the property of being able to separate the different colors locally. This vector is then added to the global luminance vector resulting in a direction that includes both global and local changes. In a region with a flat uniform color the algorithm returns the global luminance. The more color variations there are locally the more the luminance vector will be shifted to achieve the increased separation and contrast.