Android Optical Character Recognition Solution Based on TacoTaco Framework
Hang, Duy Khiem (2019)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-201904185562
https://urn.fi/URN:NBN:fi:amk-201904185562
Tiivistelmä
TacoTaco is an Android application development framework based on Model-View-Intent (MVI) model and Robert Martin's Clean Architecture. It serves as the main Android framework in the internal development processes of Wolt Enterprise Osakeyhtiö. Although it has been used extensively in the primary Android app of the company, its practicality in a generic Android application has not been thoroughly tested.
The thesis aims to provide a comprehensive analysis to ascertain if the framework can perform well against a different concrete use case. This was achieved by building an Android application using the framework as its base.
Optical Character Recognition (OCR) is a technology used to transfer text from one media to another, e.g. from paper to a digital format. Even though the technology has been around for almost a century, its recent surge in popularity stemmed the invention of smartphones. Thus, it serves as a reliable use case for the application. Because of its versatility and accuracy, the Tesseract OCR engine was chosen as the fundamental upon which the application is built.
It is concluded from the application that TacoTaco fits for general usage and that the OCR results from Tesseract are decently collected. However, more comparison with other frameworks and OCR engines, as well as in-depth performance optimization, would be a good addition to further verify their usages.
The thesis aims to provide a comprehensive analysis to ascertain if the framework can perform well against a different concrete use case. This was achieved by building an Android application using the framework as its base.
Optical Character Recognition (OCR) is a technology used to transfer text from one media to another, e.g. from paper to a digital format. Even though the technology has been around for almost a century, its recent surge in popularity stemmed the invention of smartphones. Thus, it serves as a reliable use case for the application. Because of its versatility and accuracy, the Tesseract OCR engine was chosen as the fundamental upon which the application is built.
It is concluded from the application that TacoTaco fits for general usage and that the OCR results from Tesseract are decently collected. However, more comparison with other frameworks and OCR engines, as well as in-depth performance optimization, would be a good addition to further verify their usages.