Master  Dissertation Final Defense

Discussion of a master's thesis in the College of Computer Science and Mathematics - Software Department

Entitled (Developing Reusable Handwritten Text Recognition Software Based on TensorFlow)

On Thursday 10/11/2022, the College of Computer Science and Mathematics at the University of Mosul discussed a master’s thesis for the student (Ahmed Abdulrahman Idris Alkaddo), under the supervision of Prof. Dr. Dujan Basheer Taha.

The thesis submitted by the student dealt with software for Handwritten Recognition using machine-learning techniques

This work proposes an algorithm that uses a hybrid (CNN & RNN). It will sequentially read all letterings and this is what makes it distinctive because it deals with different image dimensions (such as the word) and does not require identification of these letters that are inside the image at the training stage, because it can process this photo directly. All of these features in CRNN make it one of the best ways to read and distinguish sequential data because we will not have to hash the sequential characters, but the algorithm will do the feature extraction and distinguish the sequential characters. One of the significant portions needed to build a robust and high-accuracy model is to provide sufficient data to train this model so that it can distinguish and predict different patterns accurately.

Using this efficient method and big data, a program was built to read the image consisting of a set of lines. Where the program detects each line separately, identifies it, breaks it into words, and enters these words into the trained model to distinguish all the words in these lines in a sequential manner.


The discussion committee included the following members:

  • Dr. Shahba Ibrahim Khalil   (University of Mosul)/ Chairman
  • Associate Prof. Hanan Hamed Ali (University of Mosul)/ Member
  • Associate Prof. Dr. Meshary Ayed Askar (University of Tikrit)/ Member
  • Dr. Dujan Basheer Taha (University of Mosul) Supervisor and Member


Also Read