تفاصيل الرسالة العلمية

At present, the color image recognition system is used in many vital applications such as military, civilian and medical applications; image recognition is defined as special software that capable to identify images, such as people, faces, patterns, and fingerprint images. This study focuses on analyzing different methods that are used to extract features of a color image, color image recognition and to improve the sensitivity of features when any image processing is applied such as rotation, reshaping, and rearranging the color channels. The proposed method is based on K_means clustering algorithm which is applied on various color images, the proposed method is implemented by using a histogram to reduce input data set, then find unique features using K_means clustering algorithm, these stable features can be used as an image key (identifier), to facilitate a color image recognition after applying the image processing stage. The study introduces the method which can create a stable and unchangeable feature for a digital color image, these features will not be changed even if the image was rotated or reshaped vertically or horizontally or if the color channel rearranged in the color image, which improves the color image recognition system