Image segmentation

1. A method for classifying pixels in an image as being associated with a feature of interest, the method comprising:

(a) obtaining an input data set comprising color values for an array of pixels representing the image;

(b) segmenting, with a computer, the image based on the color values to provide a plurality of segments, each segment comprising a group of contiguous pixels with related color values;

(c) generating, with the computer, a list of all pairs of the plurality of segments which are adjacent to one another in the image;

(d) determining, with the computer, a difference in average color value between the segments comprising each pair in the list;

(e) selectively joining, with the computer, pairs of adjacent segments to form a joined segment to replace the corresponding pair of segments in the list according to at least one pre-defined joining criterion, wherein pairs of segments are considered for joining in an order based on the respective differences in average color values;

(f) classifying, with the computer, pixels as being associated with a feature of interest based on the segments following the selective joining; and

(g) providing an output data set indicating pixels classified as being associated with the feature of interest.

2. The method of claim 1 , wherein the at least one pre-defined joining criterion of step (e) is such that a pair of segments is not joined if their combined size in the image is greater than a pre-defined maximum segment size threshold.

3. The method of claim 1 , wherein the at least one pre-defined joining criterion of step (e) is such that a pair of segments is joined if at least one of the segments has a size in the image that is less than a pre-defined minimum segment size threshold.

4. The method of claim 1 , wherein the at least one pre-defined joining criterion of step (e) is such that a pair of segments is not joined if the resulting average color value for the joined segment would differ from the average color value of any of the segments from step (b) which comprises the pair of segments by more than a pre-defined maximum average color difference threshold.

5. The method of claim 1 , wherein step (e) is performed such that pairs of segments are considered for joining in an order in which the pair of segments currently having the smallest difference in average color values and which has not previously been considered for joining is the pair next considered for joining

6. The method of claim 1 , further comprising identifying a subset of the pixels potentially associated with the feature of interest based on their color values, and wherein the segmenting of step (b) is applied to pixels within this subset.

7. The method of claim 6 , further comprising adding to the subset of pixels further pixels which are not included in the subset of pixels based on their color values, but which are fully surrounded in the image by a contiguous region of pixels which are included in the subset of pixels based on their color values.

8. The method of claim 1 , wherein the segmenting of step (b) is based on a Color Structure Code (CSC) algorithm.

9. The method of claim 1 , wherein the segmentation of step (b) is performed such that a typical feature of interest in the image having an associated typical range of color values is segmented into multiple segments.

10. The method of claim 9 , wherein the number of multiple of segments is selected from the group comprising at least 3, at least 4, at least 5, at least 6, at least 7, at least 8, at least 9, and at least-10 segments.

11. The method of claim 1 , further comprising a step performed after step (e) of selectively splitting segments into pairs if a pre-defined splitting criterion is satisfied.

12. The method of claim 11 , wherein the splitting criterion is such that a segment having a kink in its boundary is split along a line joining to the kink.

13. The method of claim 1 , wherein step (f) comprises classifying those pixels within a subset of the segments resulting from the selective joining as being associated with a feature of interest, where the subset of segments corresponds to the segments which meet a pre-determined classification criterion.

14. The method of claim 13 , wherein the pre-determined classification criterion is based on comparing a geometric characteristic of the segments in the image with a corresponding pre-determined geometric characteristic threshold.

15. The method of claim 14 , wherein the geometric characteristic comprises a characteristic selected from the group comprising an areal extent, a linear dimension, and a measure of elongatedness.

16. The method of claim 13 , wherein the pre-determined classification criterion is based on comparing a color characteristic for the segments in the image with a corresponding pre-determined range of color characteristics.

17. An apparatus for classifying pixels in an image as being associated with a feature of interest, the apparatus comprising:

a source for an input data set comprising color values for an array of pixels representing the image;

and a processor configured to:

segment the image based on the color values to provide a plurality of segments comprising respective groups of contiguous pixels with related color values;

generate a list of all pairs of the plurality of segments adjacent to one another in the image;

determine a difference in average color between the segments comprising each pair in the list;

selectively join pairs of adjacent segments to form a joined segment to replace the corresponding pair of segments in the list according to at least one pre-defined joining criterion, wherein pairs of segments are considered for joining in an order based on their respective differences in average color values;

classify pixels as being associated with a feature of interest based on the segments following the selective joining; and

generate an output data set representing pixels classified as being associated with the feature of interest.

18. A non-transitory computer program product comprising a computer readable medium storing a plurality of instructions for controlling a processor to perform an operation for generating an output data set representing a classification of pixels in an image, the instructions comprising:

(a) obtaining an input data set comprising color values for an array of pixels representing the image;

(b) segmenting the image based on the color values to provide a plurality of segments, each segment comprising a group of contiguous pixels with related color values;

(c) generating a list of all pairs of the plurality of segments which are adjacent to one another in the image;

(d) determining a difference in average color value between the segments comprising each pair in the list;

(e) selectively joining pairs of adjacent segments to form a joined segment to replace the corresponding pair of segments in the list according to at least one pre-defined joining criterion, wherein pairs of segments are considered for joining in an order based on the respective differences in average color values;

(f) classifying pixels as being associated with a feature of interest based on the segments following the selective joining; and

(g) providing an output data set indicating pixels classified as being associated with the feature of interest.