Performance improvement in preprocessing phase of fingerprint recognition

Abstract

This paper introduces a new efficient algorithm for enhance the performance of preprocessing task for fingerprint recognition. It includes normalization, orientation estimation, image enhancement, binarization, and thinning process as a part of preprocessing. It introduces the improved O’Gorman filter for enhancing the image which is degraded and corrupted due to variation in skin and impression conditions as well as the gradient-based orientation estimation method is also enhanced to remove the inconsistency in ridge direction. The adaptive thresholding technique is used for binarized the image. The famous Zhang–Suen’s algorithm is enhanced for improve thinning phase. The performance is measured with different measurement standards like PSNR, MSE, and computational time as well as by quality of an image. The implementation of an algorithm is done using Java language on FVC2000 and FingerDOS database.