Method and apparatus for detecting and interpreting price label text

1. A method of price text detection by an imaging controller, comprising:

obtaining, by the imaging controller, an image of a shelf supporting labels bearing price text;

generating, by the imaging controller, a plurality of text regions containing candidate text elements from the image;

assigning, by the imaging controller, a classification to each of the text regions, selected from a price text classification and a non-price text classification for a respective text region, wherein the price text classification includes a non-numeric text element;

wherein the imaging controller, within each of a subset of the text regions having the price text classification:

detects a price text sub-region by: (a) assigning the candidate text elements within the text region to groups based on respective sizes of the candidate text elements, and (b) selecting a group having a largest area as a primary one of the groups; and

generates a price text string by applying character recognition to the price text sub-region; and

presenting, by the imaging controller, locations of the subset of text regions, in association with the corresponding price text strings.

2. The method of claim 1 , further comprising: prior to generating the text regions, identifying the candidate text elements comprises applying a blob detection operation to the image.

3. The method of claim 2 , wherein generating the text regions comprises:

for each pair of the candidate text elements, determining whether a distance between the pair is below a distance threshold.

4. The method of claim 2 , wherein generating the text regions comprises:

for each pair of the candidate text elements, determining whether a difference between a size of each of the pair is below a size threshold.

5. The method of claim 1 , wherein assigning the classification to each of the text regions comprises:

generating a feature descriptor for the text region;

providing the feature descriptor to a classifier; and

receiving the classification from the classifier.

6. The method of claim 1 , wherein detecting the price text sub-region comprises:

fitting upper and lower bounding lines to the primary group; and

extending the primary group along the bounding lines to define the price text sub-region.

7. The method of claim 6 , further comprising: prior to assigning the candidate text elements within the text region to groups, binarizing the text region.

8. The method of claim 6 , wherein at least one of the upper and lower bounding lines intersects a candidate text element in a group other than the primary group.

9. The method of claim 1 , the presenting further comprising presenting a confidence level corresponding to each price text string.

10. A server for detecting price text, comprising:

a memory storing an image of a shelf supporting labels bearing price text;

an imaging controller coupled to the memory, the imaging controller comprising:

a text region generator configured to generate a plurality of text regions containing candidate text elements from the image;

a classifier configured to assign a classification to each of the text regions, selected from a price text classification and a non-price text classification for a respective text region, wherein the price text classification includes a non-numeric text element;

a sub-region generator configured to detect a price text sub-region within each of a subset of the text regions having the price text classification by: (a) assigning the candidate text elements within the text region to groups based on respective sizes of the candidate text elements, and (b) selecting a group having a largest area as a primary one of the groups; and

an interpreter configured to generate a price text string by applying character recognition to the price text sub-region; and to present locations of the subset of text regions, in association with the corresponding price text strings.

11. The server of claim 10 , the text region generator further configured, prior to generating the text regions, to identify the candidate text elements by applying a blob detection operation to the image.

12. The server of claim 11 , the text region generator configured to generate the text regions by:

for each pair of the candidate text elements, determining whether a distance between the pair is below a distance threshold.

13. The server of claim 11 , the text region generator configured to generate the text regions by:

for each pair of the candidate text elements, determining whether a difference between a size of each of the pair is below a size threshold.

14. The server of claim 10 , the classifier configured to assign the classification to each of the text regions by:

generating a feature descriptor for the text region;

providing the feature descriptor to a classifier; and

receiving the classification from the classifier.

15. The server of claim 10 , the sub-region generator configured to detect the price text sub-region by:

fitting upper and lower bounding lines to the primary group; and

extending the primary group along the bounding lines to define the price text sub-region.

16. The server of claim 15 , the sub-region generator further configured, prior to assigning the candidate text elements within the text region to groups, to binarize the text region.

17. The server of claim 15 , wherein at least one of the upper and lower bounding lines intersects a candidate text element in a group other than the primary group.

18. The server of claim 10 , the interpreter further configured to present a confidence level corresponding to each price text string.