Live coherent image selection

1. A computer-implemented method, comprising:

defining a plurality of piecewise-constant regional cost values including a foreground cost value, a background cost value, and a bias cost value;

receiving user input defining a sample of pixels from an image defined by a raster of pixels;

classifying pixels in the raster of pixels as being foreground or background based on the sample of pixels and the plurality of piecewise-constant regional cost values; and

updating a rendering of the image on a display to depict classified foreground pixels and background pixels.

2. The method of claim 1 , wherein, while the user input is being received, the pixels in the raster of pixels are classified as being foreground or background based on the sample of pixels and the plurality of piecewise-constant regional cost values.

3. The method of claim 2 , wherein the coherent classification is based on the sample of pixels currently being received and any other samples of pixels previously received.

4. The method of claim 2 , wherein, while the user input is being received, the rendering of the image on the display is updated to depict the classified foreground pixels and background pixels.

5. The method of claim 4 , wherein the foreground cost value is associated with a foreground classification, the background cost value is associated with a background classification, and the bias cost value is associated with a bias toward the background classification or the foreground classification, and wherein classifying pixels includes coherently classifying pixels into a plurality of coherent regions, with each region classified as being foreground or background.

6. The method of claim 5 , wherein coherently classifying includes:

coherently classifying pixels of a lower resolution version of the image to generate a classification of the lower resolution version of the image;

determining a border area of the image based on the classification of the lower resolution version of the image; and

coherently classifying the border area of the image at a full resolution.

7. The method of claim 5 , wherein coherently classifying includes minimizing an objective cost function that includes regional costs and boundary costs, and minimizing the objective cost function includes incrementally solving one or more graph cut problems.

8. The method of claim 5 , further comprising updating a previous classification of pixels based on the coherent classification of pixels.

9. The method of claim 5 , wherein coherently classifying includes:

segmenting the image into a plurality of segments using a watershed or seed fill segmentation to segment the image, each segment in the plurality of segments containing a contiguous plurality of pixels; and

coherently classifying segments in the plurality of segments as being foreground or background based on the sample of pixels.

10. The method of claim 5 , wherein coherently classified foreground pixels represent one or more regions of the image that respect object boundaries in the image and do not have areas of misclassification.

11. The method of claim 1 , wherein the sample of pixels represents only foreground pixels or only background pixels.

12. The method of claim 1 , wherein the user input is received as the user draws a stroke across the image.

13. The method of claim 1 , wherein the rendering of the image on a display to depict classified foreground pixels and background pixels includes one or more of:

highlighting either the foreground or background pixels;

obfuscating either the foreground or background pixels;

applying a filter to either the foreground or background pixels;

rendering an image mask, the image mask depicting foreground or background pixels; or

rendering a border of areas of either the foreground or background pixels.

14. The method of claim 1 , wherein the image is in one of the following color spaces: RGB, CMYK, CIELAB, CIE XYZ, CIE LUV, YCC, YIQ, HSB, HSL or Grayscale.

15. A computer program product, encoded on a nontransitory computer-readable medium, operable to cause data processing apparatus to perform operations comprising:

defining a plurality of piecewise-constant regional cost values including a foreground cost value, a background cost value, and a bias cost value;

receiving user input defining a sample of pixels from an image defined by a raster of pixels;

classifying pixels in the raster of pixels as being foreground or background based on the sample of pixels and the plurality of piecewise-constant regional cost values; and

updating a rendering of the image on a display to depict classified foreground pixels and background pixels.

16. The computer program product of claim 15 , wherein, while the user input is being received, the pixels in the raster of pixels are classified as being foreground or background based on the sample of pixels and the plurality of piecewise-constant regional cost values, and wherein, while the user input is being received, the rendering of the image on the display is updated to depict the classified foreground pixels and background pixels.

17. The computer program product of claim 15 , wherein the foreground cost value is associated with a foreground classification, the background cost value is associated with a background classification, and the bias cost value is associated with a bias toward the background classification or the foreground classification, and wherein classifying pixels includes coherently classifying pixels into a plurality of coherent regions, with each region classified as being foreground or background.

18. A system comprising:

a user input device; and

data processing apparatus operable to perform operations comprising:

defining a plurality of piecewise-constant regional cost values including a foreground cost value, a background cost value, and a bias cost value;

receiving user input defining a sample of pixels from an image defined by a raster of pixels;

classifying pixels in the raster of pixels as being foreground or background based on the sample of pixels and the plurality of piecewise-constant regional cost values; and

updating a rendering of the image on a display to depict classified foreground pixels and background pixels.

19. The system of claim 18 , wherein, while the user input is being received, the pixels in the raster of pixels are classified as being foreground or background based on the sample of pixels and the plurality of piecewise-constant regional cost values, and wherein, while the user input is being received, the rendering of the image on the display is updated to depict the classified foreground pixels and background pixels.

20. The system of claim 18 , wherein the foreground cost value is associated with a foreground classification, the background cost value is associated with a background classification, and the bias cost value is associated with a bias toward the background classification or the foreground classification, and wherein classifying pixels includes coherently classifying pixels into a plurality of coherent regions, with each region classified as being foreground or background.