Simple Sequence of Procedures to Build IRS
Journal: International Journal of Computer Science and Mobile Computing - IJCSMC (Vol.13, No. 8)Publication Date: 2024-08-30
Authors : Ziad AlQadi;
Page : 41-53
Keywords : DCI; features; features database; ML; classification; ANN; Neuron; optimal ANN; training cycle;
Abstract
A simple and efficient sequence of procedures to build an image recognition system will be proposed, this system will be easily used to classify any image, with any type and size. A procedure to create the image features database will be introduced, this procedure will be based on local binary patterns method, it will decrease the features extraction time and create a unique features array for each image, and this array will be used as an identifier to recognize the associated image. The second procedure will use a feedforward artificial neural network, the architecture of this ANN will be proposed and trained, the optimal ANN which minimize the mean square error between the targets and the calculated outputs, this ANN will be saved to be used in the image classification procedure. Several images will be selected, the features of each image will be extracted and save to create the image features database. This database with the required targets will used to train the ANN; various approaches will be discussed in order to achieve the optimal ANN. The created optimal ANN then will be used as a classification tool to identify the image using its features.
Other Latest Articles
- DEVELOPMENT OF ELECTRONIC LOG BOOK SYSTEM FOR STUDENTS’ INDUSTRIAL WORK EXPERIENCE SCHEME (CASE STUDY: FEDERAL POLYTECHNIC ILE-OLUJI, ONDO STATE)
- Vehicle Tracking System Approaches: A Systematic Literature Review
- A Model for Reduction of Time and Space Complexity on Edge Devices
- Future of Work: The Impact of Intelligent Technologies for Leaders and Knowledge Workers
- A Quantitative Analysis of the Influence of Information Communication Technology (ICT) Revolution on Student Learning Experiences in Nasarawa State, Nigeria
Last modified: 2024-08-18 01:33:54