Computer system and method for offering coupons

A method of retraining a machine learning-based coupon offering system, the method comprising:

• receiving, at a computer system executing the coupon offering system, pre-purchase image data including at least one pre-purchase item and post-purchase image data including at least one post-purchase item, wherein the pre-purchase image data and the post-purchase image data are from the same purchase transaction;
• processing, by a first item detection module executed by the computer system, the pre-purchase image data to obtain first detected item data;
• processing, by a second item detection module executed by the computer system, the post-purchase image data to obtain second detected item data;
• wherein at least one of the first item detection module and the second item detection module uses a machine learning model to generate the first detected item data or the second detected item data, respectively;
• wherein at least one of the first detected item data and the second detected item data is used by a coupon matcher engine to retrieve a matched coupon from a coupon database;
• generating, by the computer system, comparison data by comparing the first detected item data and the second detected item data, the comparison data indicating one or more items that were not correctly detected by the machine learning model when generating the first detected item data or the second detected item data;
• retraining, by the computer system, the machine learning model based on the comparison data to obtain a retrained machine learning model;
• and
• deploying the retrained machine learning model to the first item detection module or the second item detection module for future processing of the pre-purchase image data or the post-purchase image data.

At least one non-transitory computer-readable storage medium storing processor executable instructions that, when executed by at least one processor, cause the at least one processor to perform a method comprising:

• receiving, at a computer system executing the coupon offering system, pre-purchase image data including at least one pre-purchase item and post-purchase image data including at least one post-purchase item, wherein the pre-purchase image data and the post-purchase image data are from the same purchase transaction;
• processing, by a first item detection module executed by the computer system, the pre-purchase image data to obtain first detected item data;
• processing, by a second item detection module executed by the computer system, the post-purchase image data to obtain second detected item data;
• wherein at least one of the first item detection module and the second item detection module uses a machine learning model to generate the first detected item data or the second detected item data, respectively;
• wherein at least one of the first detected item data and the second detected item data is used by a coupon matcher engine to retrieve a matched coupon from a coupon database;
• generating, by the computer system, comparison data by comparing the first detected item data and the second detected item data, the comparison data indicating one or more items that were not correctly detected by the machine learning model when generating the first detected item data or the second detected item data;
• training, by the computer system, the machine learning model based on the comparison data to obtain a retrained machine learning model;
• deploying the retrained machine learning model to the first item detection module or the second item detection module for future processing of the pre-purchase image data or the post-purchase image data.