Digitizing Receipts with OCR

Abstract

Receipt digitization carries numerous advantages over traditional paper receipts in terms of preservation. Saving images of receipts is a method of digitizing receipts, but managing (sorting, searching, editing, deleting, adding annotations) them is rather tedious. This paper analyzes current receipt managing applications in the market, and their approach in digitizing receipts. Other researchers' methods were also analyzed, and a pipeline to achieve receipt digitization is deduced. With a pipeline of various methods in each step, the strengths and flaws of each method were addressed. Improvements to those methods and to the overall usability of a receipt management system were proposed. The proposed methods were then developed into a usable product application.