JDP_V3_N4_RP3 Deepika Bancchor Siddharth Choubey Journal on Digital Signal Processing 2322–0368 3 4 15 19 FPI De-noising, Bi-Directional Recurrent Neural Network, PCA, Hybrid Methodology Fingerprint Images (FPI) are always prone to be corrupted by various sources of noise during the capture of an image, i.e., acquisition period. This paper studies the implementation of Pixel Component Analysis (PCA) algorithm with Bi- Directional Recurrent Neural Network (BRNN) which will effectively de-noise the FPI images. BRNN enables compression of non-reusable fingerprint image data points during PCA execution and can transform vector co-ordinates in a rational manner. The duration of execution of the operation is also significantly reduced. The output of the proposed model has showed an optimized performance for de-noising of FPI images. October - December 2015 Copyright © 2015 i-manager publications. All rights reserved. i-manager Publications http://www.imanagerpublications.com/Article.aspx?ArticleId=3707