Extraction and segmentation process for the Iraqi paper currency

Abstract

Different application like image segmentation, moving objects detection and objects tracking, required an object detection and segmentation techniques. Recently, these techniques become so important especially when only a specific part of image is important. This research paper presents an efficient algorithm that employed for objects detection and extraction process. This algorithm consists of a several steps and the validity of this algorithm is measured based on different denomination of Iraqi currency. These steps arrange as following: image conversion, deleting small and unnecessary objects from images, extraction of interest objects boundary, finally calculating the rotation angle automatically, and rotate image based on the calculated angle. The validity of algorithm measured on the seven denominations of Iraqi currency (250, 500, 1000, 5000, 10000, 25000, and 50000). This image saved in a database that contain 40 images for each denomination and the total images for all denominations are 280 images. After testing the validity of designed algorithm on all captured image, the algorithm shows high accuracy which equally to 99.6%.