A new research to identify the iris using morphological processes and neural networks presented by Engineer Ali Ali in the framework of his master's study in the Department of Mechatronic Engineering
Graduate student M. Ali obtained a master’s degree in mechatronic engineering after passing the public final defense on 7/26/2023 at 10 am in the seminar hall at Hamak College before the judging committee composed of Messrs. Dr. Iyad Hatem, Dr. Issa Al-Ghannam, Dr. Jubaili praise. The student obtained a score of 89%.
Iris recognition system depending on morphological operations, SFTA algorithm and MLP networks
The study was conducted under the supervision of Dr. Engineer Iyad Hatem from the Department of Mechatronics Engineering at Tishreen University. The following is a summary of the thesis:
This research studied the performance of the segmentation-based fractional tissue analysis (SFTA) algorithm in a human iris recognition system, and presented two new methods for segmenting restricted human iris images and a new method for extracting iris features (iris code). The SYN-IRIS-CASI database provided by the National Pattern Recognition Laboratory, Institute of Automation, and Chinese Academy of Sciences was used. The K-fold cross validation method, as well as the confusion matrix metrics, were used to evaluate the performance of the SFTA algorithm.
Results
The research proposed two new approaches to the iris recognition system: The first is for segmentation of the iris image with a total accuracy rate of 99.72% out of 10,000 iris images taken from the syn-iris-casia database, and the second is designed for the feature extraction phase with an average recognition rate of 99.78% using the training patterns and 85.32% with the test patterns using the same database.
We found that the SFTA algorithm is not recommended for iris recognition systems. The research also recommended the proposed method instead.
Photo Gallery
