Integrating Ear Anatomy with Modern Biometric Authentication System

Traditional biometric authentication systems that rely on a single method, like facial recognition, often have drawbacks. They are vulnerable to spoofing, perform poorly in different lighting, and lose accuracy if the face is partially covered. To tackle these problems, this research suggests a hybrid biometric authentication system that combines facial recognition and ear recognition for better reliability and security. The proposed system uses a modified ThickNet-based deep learning architecture with CNNs to extract ear features effectively. It employs OpenCV and DLIB for real-time, accurate facial recognition. By merging the decisions from both methods, the system improves authentication accuracy, reduces false acceptance and rejection rates, and increases resistance to spoofing attacks. Experimental results show that the hybrid model outperforms single-method approaches and offers a secure and efficient access control solution.