Classification of ear images according to person, age and gender with the local ternary pattern

The need to verify the identity of individuals is increasing day by day. Traditionally, passports, identity cards, keys are used in authentication systems. With such systems, passwords can also be used to increase security. Unfortunately, the disadvantages of such security systems include the loss, copying, and theft of the item used as an identity. Passwords can be forgotten. Such situations can endanger the person or put him in a difficult situation. Such shortcomings of traditional person recognition techniques cause major problems for everyone. Such situations push researchers to seek a solid, reliable and perfect personal description. This search pushes researchers to biometric systems. In this study, 2000 data, which are right and left ear images of 100 people, were collected. The attributes of these collected files were extracted with the Local Triple Pattern. For each image file, 1x512 vectors were produced. These processes were performed for all files and images were classified for person, age and gender with many different classification algorithms. While 90.2% accuracy rate was obtained for person recognition, 99.8% success was achieved for gender. Finally, the classification success rate was 86.1% for age.