Non-photorealistic sketching

1. A method for transforming an original image into a sketch image, comprising:

a computer to perform, the steps of:

segmenting the original image into plural regions to produce a segmented image, wherein the regions are demarcated by respective boundaries;

shrinking a boundary of at least one of the plural regions in the segmented image to produce a boundary-shrunk image; and

modifying at least one color in the boundary-shrunk image to produce the sketch image,

wherein the modifying of said at least one color comprises:

selecting candidate colors to populate the sketch image; constructing a graph based on a composition of the original image;

defining an energy function that provides an energy value associated with the graph; and

using the energy function to derive a desirable selection of the candidate colors for use in the sketch image.

2. The method of claim 1 , wherein the modifying of said at least one color manipulates colors in a color space that represents colors in a triple of: hue, chroma, and value.

3. The method of claim 1 , wherein the modifying of said at least one color relies on a color database, wherein the color database is formed by:

receiving a plurality of original images;

receiving a plurality of sketch images, wherein the sketch images represent human renditions of the plurality of respective original images; and

defining, on a region-by-region basis, how colors are mapped from the plurality of original images to the plurality of sketch images.

4. The method of claim 3 , wherein the defining is performed separately for foreground regions and background regions in the plurality of original images and the plurality of sketch images.

5. The method of claim 1 , wherein the energy function relies on at least one sketch-related rule to provide the energy value.

6. The method of claim 5 , wherein said at least one rule pertains to a background region of the sketch image, wherein said at least one rule comprises one or more of:

guideline (b1): a rendering color is dependent on not only an original color in the original image, but also a size-complexity and shape-complexity of a region;

guideline (b2): an isolated background region with a small size or complex boundary shape should be rendered with a light color;

guideline (b3): if two background regions are not adjacent but have similar color hue, these regions should be rendered with the same color; or

guideline (b4): the original intensity contrast between adjacent regions should be maintained.

7. The method of claim 5 , wherein said at least one rule pertains to a foreground region of the sketch image, wherein said at least one rule comprises one or more of:

guideline (f1): original chroma and value contrast between adjacent and homogeneous foreground regions should be maintained; or

guideline (f2): chroma and value should be increased to enhance the foreground.