Research on Fingerprint Image Differential Privacy Protection Publishing Method Based on Wavelet Transform and Singular Value Decomposition Technology

The technology of fingerprint recognition has been extensively utilized for identity authentication due to the inherent privacy concerns associated with fingerprint images. However, directly publishing fingerprint images on the Internet can result in the leakage of sensitive information. In order to protect the sensitive information in fingerprint images, this paper proposes a novel method called LSDP that combines wavelet transform and singular value decomposition while ensuring differential privacy protection. The primary focus of protecting fingerprint images involves safeguarding the matrix of two-dimensional images. To mitigate excessive noise generated by differential privacy when adding noise to the matrix of fingerprint images, this study employs wavelet transform as a solution. Additionally, it utilizes the exponential mechanism in conjunction with wavelet transform to select an appropriate threshold in order to reduce noise added by differential privacy. This approach is referred to as DWP. Furthermore, singular value decomposition is applied to the coefficients obtained from wavelet transforms. In order to reduce the effect of noise errors, perturbation noise is only added to the singular values instead of all coefficients. This proposed method is termed SDP. To enhance privacy protection during publication of fingerprint images, LSDP adds perturbation noise only to some of the singular values. Experimental results demonstrate that our proposed algorithm outperforms direct application of wavelet transform or singular value decomposition in terms of safeguarding fingerprint image.