An Authentic, Privacy-driven Traffic Offense Management System with Ear as a Biometric to Facilitate the Traffic Police on Indian Roads

Abstract

Traffic authorities throughout the world have used the online system to expedite the administration of traffic violations of services. however, maintaining these systems’ integrity and security continues to be a difficult task. The incorporation of ear biometric identification is suggested in this abstract as a unique strategy to improve the security of websites that track traffic violations. The position of the ear on the face is generally steady, and it serves us better when it is on the side of the face. Due to its discriminant qualities, ear biometrics have been used to identify identical twins, unlike in the beginning. Addition- ally, due to the consistent backdrop, picture segmentation of the ear from the facial area is less difficult than that of the face. In recent years, deep convolutional neural networks (CNNs) have made significant advancements in handling challenging visual identification issues, including noteworthy results in identifying real pictures. The use of Deep CNN in various challenging machine-learning issues has received substantial research. Deep CNN has also advanced significantly in its ability to recognize human faces. The aforementioned ac- accomplishments demonstrated that well-designed deep CNNs are useful tools for addressing these challenges. So, a reasonable person may wonder: How does CNN identify the human ear? Deep learning's powerful potential inspires us. Position fluctuation and partial occlusion, particularly ear occlusion, are the only two variations between ear recognition and facial recognition. Due to the aforementioned factors, ear identification may be a less complicated procedure.