Methods and apparatus to perform image analyses in a computing environment

1. An apparatus to classify a retail product tag, the apparatus comprising:

a feature extractor to generate a first image descriptor based on a first image of a first retail product tag corresponding to a first retailer category, the first image descriptor representative of one or more visual features of the first retail product tag;

a feature descriptor generator to generate a feature descriptor corresponding to the first retail product tag based on the first image descriptor and a first category signature corresponding to the first retailer category, wherein the feature descriptor is at least twice a bit length of the first image descriptor; and

a classifier to:

generate a first probability value corresponding to a first type of promotional product tag and a second probability value corresponding to a second type of promotional product tag based on the feature descriptor; and

determine whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag based on the first probability value and the second probability value.

2. The apparatus as defined in claim 1 , wherein the first retailer category is representative of a first retailer different from a second retailer that is represented as a second retailer category.

3. The apparatus as defined in claim 1 , wherein the image descriptor is a one-dimensional array that includes feature values for corresponding features.

4. The apparatus as defined in claim 1 , further including a retailer identifier extractor to extract retail product tag metadata from input image data.

5. The apparatus as defined in claim 1 , further including a seed generator to generate a seed for use in generating the first category signature corresponding to the first retailer category.

6. The apparatus as defined in claim 5 , wherein the seed generator is to generate the seed during a training phase, the training phase to be separate from a recognition phase, the classifier to determine, during the recognition phase, whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag.

7. The apparatus as defined in claim 6 , further including a category signature generator to generate the first category signature, the first category signature based on the seed and a plurality of image descriptors corresponding to second retail product tags of the first retailer category.

8. The apparatus as defined in claim 1 , further including a category signature mapper to store the first category signature in association with a retailer identifier corresponding to the first retailer category.

9. The apparatus of claim 1 , wherein the feature descriptor generator is to generate the feature descriptor by concatenating the first image descriptor and the first category signature.

10. An apparatus to classify a retail product tag, the apparatus comprising:

a feature extractor to generate a first image descriptor based on a first image of a first retail product tag corresponding to a first retailer category, the first image descriptor representative of one or more visual features of the first retail product tag;

a feature descriptor generator to generate a feature descriptor corresponding to the first retail product tag based on the first image descriptor and a first category signature corresponding to the first retailer category; and

a classifier to:

generate a first probability value corresponding to a first type of promotional product tag and a second probability value corresponding to a second type of promotional product tag based on the feature descriptor, wherein the first type of promotional product tag is a multi-buy promotional product tag, and the second type of promotional product tag is a price reduction promotional product tag; and

determine whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag based on the first probability value and the second probability value.

11. A non-transitory computer readable storage medium comprising instructions that, when executed, cause a processor to at least:

generate a first image descriptor based on a first image of a first retail product tag corresponding to a first retailer category, the first image descriptor representative of one or more visual features;

generate a feature descriptor corresponding to the first retail product tag based on the first image descriptor and a first category signature corresponding to the first retailer category, wherein the feature descriptor is at least twice a bit length of the first image descriptor;

generate a first probability value corresponding to a first type of promotional product tag and a second probability value corresponding to a second type of promotional product tag based on the feature descriptor; and

determine whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag based on the first probability value and the second probability value.

12. The non-transitory computer readable storage medium as defined in claim 11 , wherein the instructions, when executed, cause the processor to generate a seed for use in generating the first category signature corresponding to the first retailer category.

13. The non-transitory computer readable storage medium as defined in claim 12 , wherein the instructions, when executed, cause the processor to generate the seed during a training phase, the training phase to be separate from a recognition phase, the instructions to cause the processor to determine, during the recognition phase, whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag.

14. The non-transitory computer readable storage medium as defined in claim 13 , wherein the instructions, when executed, cause the processor to generate the first category signature based on the seed and a plurality of image descriptors corresponding to second retail product tags of the first retailer category.

15. The non-transitory computer readable storage medium as defined in claim 11 , wherein the instructions, when executed, cause the processor to store in a memory the first category signature in association with a retailer identifier corresponding to the first retailer category.

16. The non-transitory computer readable storage medium as defined in claim 11 , wherein the instructions, when executed, cause the processor to generate the feature descriptor as a one-dimensional array that includes feature values for corresponding features.

17. The non-transitory computer readable storage medium as defined in claim 11 , wherein the instructions, when executed, cause the processor to generate the feature descriptor by concatenating the first image descriptor and the first category signature.

18. A method comprising:

generating, by executing an instruction with a processor, a first image descriptor based on a first image of a first retail product tag corresponding to a first retailer category, the first image descriptor representative of one or more visual features;

generating, by executing an instruction with a processor, a feature descriptor corresponding to the first retail product tag based on the first image descriptor and a first category signature corresponding to the first retailer category, wherein the feature descriptor is at least twice a bit length of the first image descriptor;

generating, by executing an instruction with a processor, a first probability value corresponding to a first type of promotional product tag and a second probability value corresponding to a second type of promotional product tag based on the feature descriptor; and

determining, by executing an instruction with a processor, whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag based on the first probability value and the second probability value.

19. The method as defined in claim 18 , further including generating a seed for use in generating the first category signature corresponding to the first retailer category.

20. The method as defined in claim 19 , further including generating the seed during a training phase, the training phase to be separate from a recognition phase, the determining of whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag to be performed during the recognition phase.

21. The method as defined in claim 20 , further including generating the first category signature based on the seed and a plurality of image descriptors corresponding to second retail product tags of the first retailer category.

22. The method as defined in claim 18 , further including storing in a memory a first category signature in association with a retailer identifier corresponding to the first retailer category.

23. The method as defined in claim 18 , wherein the feature descriptor is a one-dimensional array that includes feature values for corresponding features.

24. The method of claim 18 , wherein the generating of the feature descriptor includes concatenating the first image descriptor and the first category signature.

25. A non-transitory computer readable storage medium comprising instructions that, when executed, cause a processor to at least:

generate a first image descriptor based on a first image of a first retail product tag corresponding to a first retailer category, the first image descriptor representative of one or more visual features;

generate a feature descriptor corresponding to the first retail product tag based on the first image descriptor and a first category signature corresponding to the first retailer category;

generate a first probability value corresponding to a first type of promotional product tag and a second probability value corresponding to a second type of promotional product tag based on the feature descriptor, wherein the first type of promotional product tag is a multi-buy promotional product tag, and the second type of promotional product tag is a price reduction promotional product tag; and

determine whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag based on the first probability value and the second probability value.

26. A method comprising:

generating, by executing an instruction with a processor, a first image descriptor based on a first image of a first retail product tag corresponding to a first retailer category, the first image descriptor representative of one or more visual features;

generating, by executing an instruction with a processor, a feature descriptor corresponding to the first retail product tag based on the first image descriptor and a first category signature corresponding to the first retailer category;

generating, by executing an instruction with a processor, a first probability value corresponding to a first type of promotional product tag and a second probability value corresponding to a second type of promotional product tag based on the feature descriptor, wherein the first type of promotional product tag is a multi-buy promotional product tag, and the second type of promotional product tag is a price reduction promotional product tag; and

determining, by executing an instruction with a processor, whether the first retail product tag corresponds to the first type of promotional product tag or the second type of promotional product tag based on the first probability value and the second probability value.