Open AccessDEEP MACHINE LEARNING TO ENHANCE ANN PERFORMANCE: FINGERPRINT CLASSIFIER CASE STUDY
Author(s) -
Mua’ad Abu-Faraj,
Khaled Aldebei,
Ziad Alqadi
Publication year - 2021
Publication title -
xi'nan jiaotong daxue xuebao
Language(s) - English
Resource type - Journals
ISSN - 0258-2724
DOI - 10.35741/issn.0258-2724.56.6.61
Subject(s) - artificial intelligence , computer science , artificial neural network , fingerprint (computing) , classifier (uml) , exploit , pattern recognition (psychology) , fingerprint recognition , flexibility (engineering) , machine learning , cluster analysis , process (computing) , mathematics , statistics , computer security , operating system
The human fingerprint is very important for human identification. Therefore, many works have been done to enhance the possibility of detecting a human fingerprint. This research article will explain how to build a high-performance fingerprint classification system. In the study presented in this research paper, the Kmeans clustering method will be used for several reasons, the most important of which is the flexibility of this method by controlling the number of values in the features vector of the fingerprint and more than one alternative can be used to form the features, as it is not sensitive to the process of rotating the fingerprint image. The key idea is to normalize the created fingerprint database and exploit the concept of deep learning to minimize the error between the calculated output and the desired target output. The artificial neural network architecture of different types of ANN will be expanded vertically and horizontally. These architectures will be trained and tested; the obtained training and testing results will be used to make some useful recommendations for using an artificial neural network as a recognition tool in the fingerprint recognition system.