An Experimental Performance Analysis on Robotics Process Automation (RPA) with Open Source OCR Engines: Microsoft Ocr and Google Tesseract OCR

Abstract

Robotic Process Automation (RPA) plays an important role in the solution for automating human digital interaction in various software domains like IT, Business Strategy and so on. Optical Character Recognition(OCR) superimposes subtitled characters on an image.Here we use two Open source OCR engines, Google Tesseract OCR-It literally makes use of the open source Tesseract OCR Engine[1]. Microsoft OCR-This is another open source OCR engine accessible in the Robotics Process Automation tool, UiPath[1]. The technique of optical character recognition (OCR) has been used to transform printed text into editable text. In various applications, OCR is a very useful and common technique. OCR precision can be dependent on algorithms for text preprocessing and segmentation. Often because of different scale, design, orientation, complex picture context, etc., it is difficult to retrieve text from the image. We begin this paper with an introduction to the Optical Character Recognition (OCR) process, in this paper, we propose a research analysis to estimate the accuracy of two open source OCR engines with Robotics Process Automation(RPA). These RPA processes are created using the UiPath tool and our results compare the OCR engines performance during the execution with the workflow and save the results to our local storage or on cloud storage if cloud support enabled and here we used local storage for purpose, and the result analysis and error analysis was evaluated by manual process which also included excel formulas to calculate the accuracy of the extracted string with the original string.