FINGERPRINT DETECTION USING DEEP LEARNING

The system allows users to upload a dataset of fingerprint images, preprocess them, and train a CNN model for live vs. fake fingerprint detection.An alternative model using a simplified VGG16-like structure is also implemented for comparison purposes.Once trained, the models can predict the authenticity of a given fingerprint image with associated confidence scores.During prediction, the system applies multiple image processing techniques such as grayscale conversion, HSV transformation, and Canny edge detection to visualize intermediate steps and aid understanding.The trained models and their performance metrics, including accuracy and loss, are stored and can be visualized using built-in plotting functions.Additionally, a comparative analysis of CNN and VGG16 performance is provided through a bar chart.Overall, this system serves as a practical tool for demonstrating how deep learning models can be used in biometric security applications to combat spoofing attacks and enhance fingerprint authentication systems.