Fingerprint Recognition Using Convolution Neural Network

Abstract - Fingerprint recognition has emerged as a highly reliable biometric technique for identity verification due to the uniqueness and permanence of fingerprint patterns. This project proposes a deep learning-based approach for fingerprint recognition using Convolutional Neural Networks (CNNs), enhanced by image inversion and data augmentation techniques. The CNN architecture is trained to automatically extract and learn robust features from fingerprint images, eliminating the need for manual feature engineering.To improve model performance and generalization, image preprocessing through inversion enhances ridge-valley contrast, making distinguishing patterns more prominent.Furthermore, various data augmentation strategies such as rotation, scaling, flipping, and noise addition are employed to expand the dataset diversity and prevent overfitting. The proposed system is evaluated on benchmark fingerprint datasets, achieving high accuracy and robustness even under noisy or partial print conditions. Key Words:Fingerprint Recognition, Convolutional Neural Network (CNN),Biometric Authentication,Image Inversion,Data Augmentation,Deep Learning,Pattern Recognition,Feature Extraction, Ridge-Valley Enhancement,Fingerprint Matching