Accurate automatic personal identification is critical in a variety of applications in our electronically interconnected society. So, this paper presents a fingerprint biometric system based on texture descriptors. For this, it requires an image of high PSNR value so that complete features must be extracted and all matching points will be obtained. For this, it uses an contrast based enhancement algorithm for improving PSNR value. We have considered the factors relating to obtaining high performance feature points detection algorithm, such as image quality, separation, image improvement and feature detection. Commonly used features for increasing fingerprint image quality are features vectors and local orientation. Accurate separation of fingerprint ridges from noisy background is necessary. A pre-processing method containing of field orientation, frequency estimation, filtering, segmentation and enhancement is performed. Image normalization is also done for equalizing the features values. Also area of interest is also found out. All simulations are done in MATLAB tool.