This research primarily focuses on fingerprint identification system. Fingerprint identification is a biometric authentication system that uses a computer database of fingerprint records to confirm an individual's identity. Here for training and testing purpose we have used custom dataset which comprises of several scanned fingerprint images. To delve deeper into this area, we have referred to previous researches intending to obtain the best results. To enhance the fingerprint image and extricate the required features here we have applied the Quaternion Kalman filter. This filter is based upon the conception of hyper-complex numbers. In our study we have used the loss function of quaternion filter for obtaining the best outcomes. Subsequently, in order to classify the extracted features Multilayer Perceptron (MLP) classification approach is suggested in research. MLP is an Artificial Neural Network (ANN) which is used to solve supervised learning problems. In the suggested methodology, MLP is implemented for classifying the enhanced features of the original fingerprint image and the recognized one. Thereafter, Fingerprint matching is performed using quaternion algorithm which is based upon neural networks. It helps to get better visualization of an image. Proposed methodology leads us towards the finest results.

Fingerprint Recognition, quaternion, MLP, PCA