Simple Sequence of Procedures to Build IRS

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.