Combining Gabor filtering and classification dictionaries learning for fingerprint enhancement

Abstract

This study presents a new method for enhancing fingerprint image. The process of the enhancement is divided into two phases: fingerprint is first enhanced using Gabor filtering and then the enhanced fingerprint can be further enhanced by using sparse representation with the priori information of ridge pattern based on classification dictionaries learning. In the second stage, first, the orientations of fingerprint patches are estimated by the weighted linear projection analysis and the quality of patches are evaluated by the coherence of point orientations. Second, the training patches are classified into eight groups based on their own orientations, and the training samples of each class are selected from candidate patches by their own quality. The corresponding classification dictionaries are learned in frequency domain. Finally, the fingerprint image is enhanced based on spectra diffusion by using classification dictionaries learning. The experiments are carried out using various fingerprint enhancement methods. The experiments show that the proposed method achieves better results in comparison with other methods, and can significantly improve the performance of automatic fingerprint identification system.

Fingerprint enhancement using multi‐scale classification dictionaries with reduced dimensionality

“…To validate our proposed method on fingerprint enhancement performance, the performance of the proposed algorithm is an evaluation based on the public competition fingerprint databases FVC 2000 and FVC 2004. The comparative experimental results demonstrate that the proposed method is more effective and efficient in fingerprint image enhancement than the existing methods such as Gabor filter method [6], Chikkerur's STFT method [15], Sutthiwichaiporn's ABSF method [16], Ding's ACBF [17] method and Ding's CDL method [31].…”

“…The computational complexity of each enhancement algorithm can be measured by its average execution time. The proposed method and the similar CDL method proposed in [31] are employed to enhance the fingerprints in FVC2000 DB1_B, and the corresponding execution times are shown in Fig. 7.…”

“…The original fingerprint is shown in row 1 column 1. Correspondingly, the enhanced fingerprints by Gabor filtering method [6], STFT method [15], ABSF method [16], ACBF method [17], CDL method [31] and proposed method are shown from row 1 column 2 to row 2 column 3 Fig. 10 Fingerprints enhanced by various methods, the original fingerprint image comes from FVC2004 DB2 16_8.…”

“…It has made great achievements in face recognition [20,21], image denoising [22,23], image superresolution [24,25] and image fusion [26,27]. The relatively less active research on fingerprint image sparse representation also starts to attract the researcher's attention [28][29][30][31]. The focuses of these researches are mainly on exploring new and efficient methods to improve the performance of the AFIS, and some achievements and progresses have been made recently in this area.…”

“…To overcome the above-mentioned limitations and flaws in [31], in this research, we propose a robust fingerprint enhancement via sparse representation over learned multi-scale classification dictionaries with reduced dimensionality. The anti-noise ability of fingerprint patch enhancement varies in relation to the patch size.…”