Fingerprint generation and authentication though Adaptive convolution generative adversarial network (ADCGAN)

Fingerprints are crucial in identification of humans. The uniqueness of finger prints makes it an interesting subject. Fingerprints are termed as a technique used to define, assess, and quantify a person's physical and behavioral property. Deep learning has made its application in all the major fields such as natural language processing, computer vision and speech processing. Deep learning has also found its application in the important subject of fingerprint synthesis and biometric. The ever-growing complexity of fingerprint authentication issues, from cellphone authentication to airport security systems, seems to be best handled by these models. In recent years, deep learning-based models have been used more and more to raise the accuracy of various fingerprint recognition systems. The persuasive capacity of Generative Adversarial Networks (GANs) to generate believable instances can be credibly taken from an existing distribution of samples. GAN exhibits exceptional performance on data generation-based tasks and also encourages study in privacy and security. In this work, using Adaptive Deep Convolution Generative Adversarial Networks (ADCGAN), we develop a model that generates and authenticate the fingerprints. A Socofing dataset was trained on ADGAN model. The model gave 92% accuracy. The conduct of fingerprint research has been made possible due to ADGAN, without restrictions related to the confidential nature of biometric data.