System and method for retail store promotional price tag detection and maintenance via heuristic classifiers

Patent metadata

Field	Value
Publication number	US-2016260051-A1
Application number	US-201514637830-A
Country	US
Kind code	A1
Filing date	Mar 4, 2015
Priority date	Mar 4, 2015
Publication date	Sep 8, 2016
Grant date	—

Abstract

A system to detect and maintain retail store promotional price tags (PPTs) includes a heuristic PPT description extractor module, a heuristic rule deriver module, a store shelf image acquisition system, a barcode locator and recognizer module, and a heuristic PPT classifier module. The heuristic PPT description extractor module extracts heuristic descriptions of PPTs. The heuristic rule deriver module derives a set of heuristic parameters for the PPTs. The barcode locator and recognizer module analyzes images acquired by the store shelf image acquisition system to localize and recognize barcodes. The heuristic PPT description classifier module extracts heuristic attributes from the images acquired by the store shelf image acquisition system using the set of PPT parameters supplied by the heuristic rule deriver module.

First claim

What is claimed is:

1 . A system to detect and maintain retail store promotional price tags comprising:
a heuristic promotional price tag description extractor module adapted to extract heuristic descriptions of promotional price tags; a heuristic rule deriver module in communication with the heuristic promotional price tag description extractor module, the heuristic rule deriver module being adapted to derive a set of promotional price tag parameters; a store shelf image acquisition system adapted to acquire shelf images in a retail store; a barcode locator and recognizer module in communication with the store shelf image acquisition system, the barcode locator and recognizer module being adapted to analyze images acquired by the store shelf image acquisition system to localize barcodes; and a heuristic promotional price tag classifier module in communication with the barcode locator and recognizer module and the heuristic rule deriver module, the heuristic promotional price tag description classifier module being adapted to extract heuristic attributes from the images acquired by the store shelf image acquisition system using the set of promotional price tag parameters supplied by the heuristic rule deriver module.

2 . A method of detecting and maintaining retail store promotional price tags using heuristic classifiers comprises:
extracting heuristic descriptions from at least one exemplar promotional price tag having a barcode; deriving heuristic parameters from the exemplar promotional price tag heuristic descriptions; acquiring a plurality of shelf images at a retail store; analyzing the acquired shelf images to localize barcodes; extracting heuristic attributes from the acquired shelf images; determining whether a localized barcode is related to a promotional price tag; and reporting missing or invalid promotional price tags.

3 . The method of claim 2 wherein extracting heuristic descriptions from at least one exemplar promotional price comprises:
displaying at least one selected exemplar promotional price tag from the retail store; and
prompting an operator to identify at least two heuristic elements of the selected exemplar promotional price tag, a one of the at least two heuristic elements being a barcode element, other of the at least two heuristic elements being non-barcode elements.

4 . The method of claim 3 wherein a plurality of exemplar promotional price tags are displayed and extracting heuristic descriptions from at least one exemplar promotional price also comprises:
1) prompting an operator to select one of the displayed exemplar promotional price tags;
2) prompting the operator to identify at least two heuristic elements of the selected exemplar promotional price tag, a one of the at least two heuristic elements being a barcode element, other of the at least two heuristic elements being non-barcode elements; and
3) repeating steps 1 and 2 until each of the displayed exemplar promotional price tags has been selected.

5 . The method of claim 4 wherein extracting heuristic descriptions from at least one exemplar promotional price may comprises prompting the operator to confirm that the exemplar promotional price tags not selected in step 3 include all of the heuristic elements identified in step 2 .

6 . The method of claim 4 wherein extracting heuristic descriptions from at least one exemplar promotional price also comprises combining and merging the identified heuristic elements into a single set of promotional price tag descriptions.

7 . The method of claim 3 wherein deriving heuristic parameters from the exemplar promotional price tag heuristic descriptions comprises:
measuring the size of the barcode of the selected exemplar promotional price tag in pixels;
computing baseline scores for the non-barcode heuristic elements;
aggregating the baseline scores of the non-barcode heuristic elements to produce an overall baseline score; and
publishing the overall baseline score.

8 . The method of claim 7 wherein computing baseline scores for the non-barcode heuristic elements comprises:
detecting attributes of the non-barcode heuristic element;
measuring the detected attributes of the non-barcode heuristic element; and
determining a baseline score for the non-barcode heuristic element based on the measured detected attributes.

9 . The method of claim 8 wherein if the non-barcode heuristic element is color patch, detecting attributes of the non-barcode heuristic element includes performing a color patch detection on the selected exemplar promotional price tag, and measuring the detected attributes includes at least one of measuring size, measuring aspect-ratio, measuring fill-factor, or measuring position relative to the barcode.

10 . The method of claim 8 wherein if the non-barcode heuristic element is number, detecting attributes of the non-barcode heuristic element includes performing a digit detection on the selected exemplar promotional price tag, and measuring the detected attributes includes at least one of measuring size, or measuring position relative to the barcode.

11 . The method of claim 8 wherein if the non-barcode heuristic element is text, detecting attributes of the non-barcode heuristic element includes performing a text detection on the selected exemplar promotional price tag, and measuring the detected attributes includes at least one of measuring size, measuring similarity to an ideal text, or measuring position relative to the barcode.

12 . The method of claim 8 wherein if the non-barcode heuristic element is logo, detecting attributes of the non-barcode heuristic element includes performing a logo detection on the selected exemplar promotional price tag, and measuring the detected attributes includes at least one of measuring size, measuring similarity to an ideal logo, or measuring position relative to the barcode.

13 . The method of claim 7 wherein analyzing the acquired shelf images comprises:
selecting one of the acquired shelf images;
analyzing the selected acquired shelf image; and
determining if a barcode is present in the selected acquired shelf image by localizing the barcode.

14 . The method of claim 13 wherein extracting heuristic attributes from the acquired shelf images comprises:
extracting heuristic attributes from at least one portion of the selected acquired shelf image; and
determining whether or not the localized barcode is part of a promotional price tag.

15 . The method of claim 14 wherein extracting heuristic attributes from the acquired shelf images comprises:
producing a sub-image from the acquired shelf image;
comparing heuristic elements of the sub-image to heuristic elements of the exemplar promotional price tags; and
determining if the sub-image includes a promotional price tag.

16 . The method of claim 15 wherein producing a sub-image from the acquired shelf image includes:
measuring a size of the selected barcode in pixels; and
cropping a portion of the acquired image to produce a sub-image.

17 . The method of claim 15 wherein extracting heuristic attributes from the acquired shelf images also comprises:
measuring the size of the barcode of the sub-image in pixels;
computing sample scores for the non-barcode heuristic elements;
aggregating the sample scores of the non-barcode heuristic elements to produce an overall sample score;
publishing the overall sample score; and
comparing the overall sample score to the overall baseline score to determine if the sub-image includes a promotional price tag.

Assignees

• Xerox Corp

Inventors

• Wu Wencheng
• Venable Dennis L
• Wade Thomas F
• Paul Peter
• Cote Adrien P

Classifications

• G06T11/60

  Creating or editing images; Combining images with text · CPC title

• G06Q10/087 Primary

  Inventory or stock management, e.g. order filling, procurement or balancing against orders · CPC title

• G06K9/4604

  Physics · mapped topic