Optical Character Recognition (OCR) of The Multi Photos

Abstract

This study shows how successful Optical Character Recognition (OCR) systems are when they control multiple photos instead of a single photo. An exemplary android application was developed to set up this system. This application checks whether the ingredients that were previously added to the system are in the product after the photos of the ingredient list are taken. To better understand the system, theoretical aspects were defined, and the system was tested with a single photo input and multiple photo input. In this experimental study, it was explained how the system was designed and how it was implemented. After these stages, the system was tested and the success rates of Optical Character Recognition (OCR) with a single photo input and multiple photo input were determined. In order to obtain the most consistent and accurate results, different experimental scenarios were prepared for both processes. As a result of these tests, Optical Character Recognition of multi photo is more successful than OCR of single photo.