Accurate Fingerprint Recognition and Gender Classification Using Inception V3

Aim: This study aims to develop a secure and accurate Fingerprint Recognition (FR) system by using an Inception V3 model to classify gender from fingerprint images. The system’s predictive accuracy is evaluated and compared with that of a standalone CNN-based model. The Inception V3 model achieves an accuracy of 97%, significantly surpassing the CNN model’s accuracy of 93%. Materials and Methods: In this research, there are two groups. Group 1 is the CNN-based Fingerprint Recognition system that was tested using 20 sample images with varied accuracy. Group 2 has the proposed Inception V3 model that utilizes advanced feature extraction and classification techniques for optimal predictive accuracy. Result: The CNN model results show an accuracy of 82.50% to 87.80%, while the Inception V3 model performed better with an accuracy of 92.20% to 96.80% at similar testing conditions. Parameters for testing include a maximum improvement of 6.50% and a minimum of 5.50% with a step size of 0.20 and a significance level of p < 0.05. Conclusion: From this work, it appears that the Inception V3 model is much more accurate as compared to the CNN model.