You are cordially invited by the COE Dept., to attend a PhD proposal presentation, on the above given title, by Mr. Dhia Musleh, PhD Student (lecturer-B), ICS Dept., on Wednesday, May 22nd, 2013, at 01:00 P.M., in Building 22, Room 130.
Abstract: Nowadays, touch screen, pen-based computers/smartphones have been used widely. Online text forms a natural representation for inputting data to handheld devices and smart phones. Automatic online Arabic text recognition has many applications in data entry for these devices. Developing a system for automatic recognition of online Arabic text will provide a quick and natural way of communication between computers, digital assistants, smart phones and human.
In this thesis, we will conduct research on automatic recognition of online handwritten Arabic text. This is expected to result in developing the theory and algorithms of online Arabic text recognition. Furthermore, a prototype system will be built to evaluate the proposed work of this thesis.
Online Arabic handwritten recognition has been less researched compared to offline Arabic handwritten recognition. The reasons for this may be due to the challenges related to online Arabic handwriting recognition systems. Some of these challenges include the need for special hardware; techniques of other languages may not work for online Arabic text recognition.
In this thesis, we will apply structural-based techniques to Arabic online text recognition since they proved to have satisfying results with offline Arabic handwritten text. In addition, we will adapt the fuzzy attributed turning functions proposed by (T. Parvez & Sabri 2013) and extended by (Halawani 2013) to address online Arabic text. We will develop the previous work to be more robust by having improved fuzzy models and similarities measures. We expect to produce novel features, fuzzy models and similarity techniques, algorithms, etc. Machine learning techniques will be used to achieve these results. The results of the proposed work will be compared with the results of other published research to evaluate the performance of the proposed work.