Study of D-CHK
Title of the academic study:
Performance Boosting of Image Matching based Iris Recognition Systems using Deformable Circular Hollow Kernels and Uniform Histogram Fusion Images
Harisu Abdullahi SHEHU, Ibrahim Furkan INCE, Faruk BULUT
The study is submitted to a journal. It is under review. If it is accepted for publication, later, give might a reference of our study in your paper as:
Harisu Abdullahi SHEHU, Ibrahim Furkan INCE, Faruk BULUT, Performance Boosting of Image Matching based Iris Recognition Systems using Deformable Circular Hollow Kernels and Uniform Histogram Fusion Images, 2022.
Web link of the article is here.
Java Codes
The Processing Java codes can be downloaded from here. They are in a zipped folder. Unzip them first. It will be better to use the Processing environment rather than NetBeans or Eclipse.
Datasets
Be careful that the dataset folder is empty. Please download the MMU1 and MMU2 datasets from the related URLs. In the article its location is defined. Replace them in the related folder before running. After running, you will see the outputs in the sub folders. Then you can start the training operation in Python.
You can also test another datasets.
Python Codes
The Python codes for the Deep Learning process can be also downloaded from here.
You can run these codes on your computer and receive the same experimental results.
Note that please
This study actually is a progressive form of a previous study in the Oxford Computer Journal:
Ince, Ibrahim Furkan, Yusuf Sait Erdem, Faruk Bulut, and Md Haidar Sharif. "A low-cost pupil center localization algorithm based on maximized integral voting of circular hollow kernels." The Computer Journal 62, no. 7 (2019): 1001-1015.
URL: https://academic.oup.com/comjnl/article-abstract/62/7/1001/5116110?login=false