Method and system for detecting and extracting price region from digital flyers and promotions

A system for detecting and extracting price region from digital flyers and promotions, comprising:

• a memory storing instructions;
• one or more communication interfaces;
• and
• one or more hardware processors coupled to the memory via the one or more communication interfaces, wherein the one or more hardware processors are configured by the instructions to:
• acquire, a set of digital flyers and promotions as input images from external sources;
• detect, one or more text regions comprising a price information from each input image from the input images by creating a bounding box around the detected texts including individual text character and broken words in each input image and merging adjacent text regions bounding boxes to form text cluster and filtering the merged text regions when length to width ratio of the bounding box lies within a predefined range, wherein the price information includes different representation of values which varies with font, color, textual information and placement of the textual information;
• convert, each text region into a two-color text comprising of a set of white pixels and a set of black pixels by, (i) changing each text region into a gray scale image and selecting top ranked two peaks from a histogram comprising one or more text regions, (ii) calculating a mean value of boundary pixels of the gray scale image, (iii) calculating a mean range based on the peaks closer to the boundary pixel mean value and the peaks farther to the boundary pixel mean value, and (iv) converting one or more pixels in black color which lies within the mean range around peak closer to the boundary pixel mean value and the one or more pixels into white color which falls out of the boundary pixel mean value, wherein converting the each text region into the two-color text for handling color variations in the input images;
• detect, the price region from the two-color text by using a price tag flag status;
• and
• extract, the price from the price region of each input image by using a price tag flag status, wherein values of the price include an irregular font size.

One or more non-transitory machine-readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause:

• acquiring a set of digital flyers and promotions as input images from external sources;
• detecting one or more text regions comprising a price information from each input image from the input images by creating a bounding box around the detected texts including individual text character and broken words in each input image and merging adjacent text regions bounding boxes to form text cluster and filtering the merged text regions when length to width ratio of the bounding box lies within a predefined range, wherein the price information includes different representation of values which varies with font, color, textual information and placement of the textual information;
• converting each text region into a two-color text comprising of a set of white pixels and a set of black pixels by, (i) changing each text region into a gray scale image and selecting top ranked two peaks from a histogram comprising one or more text regions, (ii) calculating a mean value of boundary pixels of the gray scale image, (iii) calculating a mean range based on the peaks closer to the boundary pixel mean value and the peaks farther to the boundary pixel mean value, and (iv) converting one or more pixels in black color which lies within the mean range around peak closer to the boundary pixel mean value and the one or more pixels into white color which falls out of the boundary pixel mean value, wherein converting the each text region into the two-color text for handling color variations in the input images;
• detecting the price region from the two-color text by using a price tag flag status;
• and
• extracting the price from the price region of each input image by using a price tag flag status, wherein values of the price include an irregular font size.