LSTM and CNN hybrid model for enhanced fingerprint recognition

Abstract

The paper introduces an advanced hybrid model combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to enhance fingerprint detection and matching within large datasets. The CNN component is employed for feature extraction and learning from image data, while the LSTM component is utilized for sequence prediction in temporal series, yielding optimal results compared to existing methods based on specific criteria. This hybrid approach achieves a fingerprint recognition accuracy of 99.85 %. The proposed method effectively reduces errors in recognition and false rejection rates in fingerprint recognition systems, thereby improving overall usability and security. The integration of CNN and LSTM in fingerprint recognition signifies a substantial advancement in biometric authentication technology, with potential applications in law enforcement, border security, and access control systems.