In today's digital world, security and protection play an important role in the daily use of computer systems. Current security levels can be cracked by someone at any time. A biometrically oriented security system is expected to meet users' requirements, such as lower error costs, high security levels, the ability to detect forgery, and so on. Especially recently, an automated biometric recognition system has a wide range of applications, consisting of automated recognition and information capture (OCR), automatic security verification, confirmation of private recognition to prevent information disclosure or recognition fraud, and so on (RFID). These systems require high accuracy and ease of use. In this paper, an efficient ear-focused recognition technique using Convolutional Neural Network (CNN) and Shape Mapping technique (CCM) is proposed. It is a non-intrusive method, and connections are probably among the most common biometric techniques used by people to recognize others. There are numerous advantages to using the ear as a source of information for human recognition. A biometric ear system consists of exploration of the ear and recognition of the ear. In the present work, the researchers used the K-nearest neighbor (KNN) method, which cannot cope with a large number of data sets and tries to detect the differences in each dimension. The goal of the recommended system is to improve the security and safety level using the Convolutional Neural Network (CNN). The experiments have shown that the CNN we developed achieves higher accuracy compared to existing systems.

OCR, RFID, CNN, KNN, CCM.