Project Home
PDF access

Authors

Ragaa Moustafa, Farida Hesham, Samiha Hussein, Badr Amr

Supervised by: Dr. Taraggy Ghanim, Eng. Nada Ayman, Eng. Samira Refaat

Publishing Date

November 8, 2021

Abstract

Hieroglyphs are symbol-based scripts, and each symbol can have various definitions. Therefore, understanding Hieroglyphs and having an accessible tool that accurately translates them to a commonly used language, such as English or Arabic, is quite the challenge. Hence, the idea of Scriba was brought about. Scriba is a mobile application that aims to resolve the issue of acknowledging the Hieroglyphs by providing an accessible gadget that scans and segments pictures of Hieroglyphic text using deep learning then translating it to a comprehensible language using a pre-made API. This document briefly discusses the history of Hieroglyphs and the academic and business motivation behind the project. Furthermore, it proposes the initial requirements and objectives of the mobile application, the similar systems, the proof of concept, and the project’s deliverables.

1.1 Background

Hieroglyphic writing is a system that employs characters in the form of pictures. It originated around 3100 B.C. and the last hieroglyphic carving was written in Egypt 3500 years later, in the 5th century A.D. The word “hieroglyphic” is Greek for “sacred carving”, and it was given this name since it was only carved on Egyptian monuments (i.e. tombs, temples, etc..). There were other two types of Egyptian writing called hieratic and demotic. Hieratic scripts were mainly found on papyrus or on any other smooth surface; While demotic scripts were used for ordinary documents.

Moreover, in the 2nd and 3rd centuries CE, these scripts were altered by Egypt’s conversion to Christianity. The new religion was against the Egyptian polytheism, thus the Egyptian Christians (called Coptic) started using the Greek alphabet for writing their spoken Egyptian language. Furthermore, they enlarged the Greek alphabet with seven extra letters for Egyptian sounds that do not exist in Greek. This tuned up language still exists till this day and it is called “Coptic Language”, which is used in orthodox church hymns and chants. In July 1799, Jean-Francois Champollion decoded the Hieroglyphs from the Rosetta stone and discovered that the signs were phonetic too. Therefore, Egyptologists started knowing more about ancient Egyptians and their culture by understanding the writings in temples and pyramids.

1.2 Motivation

Academic: The academic motivation is to recognize hieroglyphics using a lightweight CNN, such as ShuffleNet. Lightweight CNNs have fewer parameters than the parameters used in traditional CNN models. Therefore, they are more suited for smartphones and embedded systems. When lightweight CNNs are used instead of traditional CNNs, they can achieve the same -or better- accuracy and results.


Business: Tourism is one of Egypt’s most important sources of income; many people work in this industry, including tour guides, buyers, and tourism companies, which benefit the country’s economy. In 2019, the number of tourists visiting Egypt was 13.1 million tourists. The Egyptian ministry of education recently announced that the ancient Egyptian language will be taught in schools to educate kids about their ancient history and raise awareness. Thus, parents will use the application in the future to understand the language and teach it to their children. Furthermore, tourists will prefer reading the hieroglyphs carved in the temples and understanding more about the stories behind them to listening to unreliable tour guides describe it.

1.3 Problem Statement

There are multiple challenges in developing this application. For instance, searching for the optimal lightweight CNN to be suitable for mobile phones like ShuffleNet and MobileNet. Moreover, since the hieroglyphic writing system does not have a specific direction, it is challenging to identify whether the script is written from right-to-left (RTL) or left-to-right (LTR). In addition, Hieroglyphs can be written in rows or columns, therefore the application has to be able to recognize not only the direction of the engravings but also the context in which they are written. Moreover, the meaning of some signs can change depending on the context or the phrase. For instance, the sun disk glyph can mean both “sun” or “Ra” (the sun God). Another issue is whether the application will work offline or online. Some historical sites do not have access to the internet, such as the pyramids. Hence, an online application would not be ideal for them. That said, the performance of an offline application would not be ideal either. These challenges, as well as others, will be discussed later on.