Multibiometric classification for people based on artificial bee colony method and decision tree

Scientists confirm that a person's ear-identification system may one day be used to detect criminals and wanted persons, and may even be used to lock and unlock phones. Because of the unique privacy of the ear in the human body, it can be used to identify a person, just as it happens with a fingerprint that is unique to every person. Although the ear grows and differs in shape with age, it retains signs in terms of external appearance, which can be used in Getting to know people. However, some reasons still make identifying a person's identity from their ears difficult, such as the ears being covered with hair for example, or the necessity of photographing the ear from several different angles for different identification documents. So it would be difficult to use an ear image to identify people on a large scale, as in a fingerprint or iris. Therefore, two types of data were adopted. If the ear was not recognized, the geometry of the finger was introduced, which is also considered one of the strong points because it is not erasable and is prominent. In this paper, two datasets were used. In the event that a person is not identified by a fingerprint, it may be another fingerprint that can identify him. A real dataset was built, one for ear and the other for single finger geometry. It became about forty different people in the age group, and the images were processed by converting them to grayscale images and deletion of noise by bilateral filter and uniformity of illumination using Histogram equalization. The distinctive features of the image were extracted through the FAST method, but it was found that some features are not desirable and we need to use only strong features, so the features were selected, using one of the artificial intelligence methods, the bee colony (ABC) method, which is an important algorithm because it is inspired by the intelligent behavior of bees from their research About food and their news. the rest of the living things. Workaround the food from the dance that was used to obtain the strong attribute to facilitate the classification process, which was used as a decision tree, by training 70% dataset and testing 30%, where the results reached the accuracy of the system reaches 96.4%.