Ensaf H. Mohamed

Work place: Computer Science Department, Faculty of Computers and Information, Helwan University, Egypt

E-mail: ensaf_hussein@fci.helwan.edu.eg


Research Interests: Natural Language Processing, Image Processing, Medical Image Computing, Data Mining


Ensaf H. Mohamed received her Ph.D. in computer science, faculty of computers and information systems, Helwan University, Cairo, Egypt, 2013. Her recent research focuses on Medical Image Analysis, Machine Learning, Natural Language Processing, Text Mining, and Social Media Mining. Currently, she is an assistant professor, faculty of computers and information, Helwan University, Cairo, Egypt.

Author Articles
Automating Text Simplification Using Pictographs for People with Language Deficits

By Mai Farag Imam Amal Elsayed Aboutabl Ensaf H. Mohamed

DOI: https://doi.org/10.5815/ijitcs.2019.07.04, Pub. Date: 8 Jul. 2019

Automating text simplification is a challenging research area due to the compound structures present in natural languages. Social involvement of people with language deficits can be enhanced by providing them with means to communicate with the outside world, for instance using the internet independently. Using pictographs instead of text is one of such means. This paper presents a system which performs text simplification by translating text into pictographs. The proposed system consists of a set of phases. First, a simple summarization technique is used to decrease the number of sentences before converting them to pictures. Then, text preprocessing is performed including processes such as tokenization and lemmatization. The resulting text goes through a spelling checker followed by a word sense disambiguation algorithm to find words which are most suitable to the context in order to increase the accuracy of the result. Clearly, using WSD improves the results. Furthermore, when support vector machine is used for WSD, the system yields the best results. Finally, the text is translated into a list of images. For testing and evaluation purposes, a test corpus of 37 Basic English sentences has been manually constructed. Experiments are conducted by presenting the list of generated images to ten normal children who are asked to reproduce the input sentences based on the pictographs. The reproduced sentences are evaluated using precision, recall, and F-Score. Results show that the proposed system enhances pictograph understanding and succeeds to convert text to pictograph with precision, recall and F-score of over 90% when SVM is used for word sense disambiguation, also all these techniques are not combined together before which increases the accuracy of the system over all other studies.

[...] Read more.
Other Articles