The Computation On Fingerprint Direction Field Based On Diffusion Weight

Fingerprint recognition is a popular research topic in biometrics.At present,fingerprint identification technology is mainly used in access control,attendance systems,information security and public security case.The calculation of the fingerprint direction field is one of the most important algorithms in the fingerprint recognition algorithm,which affects the subsequent fingerprint enhancement and feature extraction.In order to accurately deal with low quality fingerprint texture features,this article starts from the direction field and studies several techniques in the fingerprint processing process:To solve the problem of low reliability of the fingerprint block in the noise region for the traditional voting method,this paper proposes a method for finding the direction field based on the diffusion of direction weights.Firstly,the square direction gradient method is used to extract the rough direction of the fingerprint field,and then the quality of the fingerprint is evaluated by using gradient consistency based on the weights.The quality block is divided into four levels.Subsequently use the neighborhood discrimination method is used to correct the misjudged fingerprint block.After correcting the misjudged quality block,the number of evaluated blocks in the neighboring block is counted,and each block with the same quality score is divided into three priority levels.Finally,the fingerprint direction field is iteratively calculated by the quality and priority.In the iteration process,different weight values are assigned according to the square gradient of the neighborhood block of the block to be evaluated.After the iteration of the evaluation block is complete,the final direction field is obtained.The experimental results show that the method proposed in this paper can improve the reliability of the low-quality region through mass diffusion,thereby improving the accuracy of the voting method in the regions of poor reliability.To solve the problem that the traditional STFT(Short Time Fourier Transform)enhancement is prone to error field in the direction of the noise region,this paper improves the traditional STFT enhancement algorithm based on the direction obtained by the direction weighted diffusion method.First,STFT analysis is applied to determine the frequency of each block and smooth the frequency.Then use the direction field of this paper is used instead of STFT to analyze the direction field obtained.Finally,the image is filtered based on the direction and frequency information.After the inverse Fourier transform,the final enhanced image is obtained.The experimental results show that the improved method in this paper has a good enhancement effect.In order to verify the performance of the proposed algorithm,we use it to perform statistical analysis on the large-scale samples of FVC2002,FVC2004 and FVC2006 fingerprint image database.The true positive rate of statistical singularity is 90%,and the total error rate of minutiae point is 22.37%.It shows that this algorithm can reconstruct the direction field of low quality fingerprint accurately.