Arificial Intelligence

The artificial neural networks (ANNs) are used to determine fingerprint patterns to obtain high fidelity matches even with poor quality prints, expediting the candidate searching in a huge database. Several neural networks, including Probabilistic Neural Network, Multi-Layer Perceptron Neural Network, and Convolution Neural Network will be compared in matching the high ranking reference candidates with the field collected latent images. A user friendly biometric image processing software will be developed for fingerprint identification, where fingerprints are featured by the location and orientation of various minutiae points. This software contains algorithms used for image conversion, feature extraction and comparison. To validate the experiments result both the algorithm MINDTCT and the proposed ANN algorithms will be compared in matching velocity and identification accuracy.