{Reliable extraction of features from poor quality prints is the most challenging problem faced in the area of fingerprint recognition. We adopted an algorithm for fingerprint image enhancement based on Short Time Fourier Transform (STFT). STFT is a well known time – frequency analysis technique to analyze non–stationary signals. The existing fingerprint feature extracting algorithms based on predominantly local landmarks (e.g., minutiae–based) or exclusively global information. Our representing fingerprint texture that relies on extracting one (or more) invariant points of reference of the fingerprint texture based on an analysis of its orientation field. A predetermined region of interest around the reference point is circular tessellated into cells. Each cell is then examined for the information in one or more different, orientation specific, spatial frequency channels. An ordered enumeration of the features thus extracted from each cell is used as the representation of the fingerprint. Thus, the representation elements capture the local information and ordered enumeration of the tessellation captures the invariant global relationship among the local patterns. The local discriminatory information in the sector, decomposed by Gabor filter bank and compute the average absolute deviation from the mean of gray value intensities is called the Finger Code},