Facial Image Publication with Differential Privacy
Zhang Xiaojian,Fu Congcong,Meng Xiaofeng(School of Computer Information Engineering,Henan University of Economics and Law;School of Informatica,Renmin University of China)
Objective Facial image publication in a direct way may lead to privacy leakage, because facial images are inherently sensitive. To protect the private information in facial image, this paper proposes an efficient algorithm based on Fourier transform combined with differential privacy. Method First, this algorithm employs the real-valued matrix to model facial image, in which each cell corresponds to each pixel point of image. After that, this algorithm relies on Fourier transform based on the matrix to extract the Fourier coefficients, and then uses the Laplace mechanism to inject noise into each coefficient to ensure differential privacy. Finally, this algorithm uses Fourier inverse transform to reconstruct the noisy facial image. However, in this process, we encounter two sources of errors: 1) the Laplace error (LE) due to Laplace noise injected, and 2) the reconstruction error (RE) caused by lossy compression. The trade-off between the LE and the AE is vital to the final accuracy of a sanitized facial image. To address this problem, this algorithm samples k elements in different candidate coefficient set via Exponential mechanism. For the samples, we add the Laplace noise to meet differential privacy. Result SVM classification on four real facial image datasets show that our proposed algorithm significantly outperforms existing solutions in terms of precision, recall, and F1-score. Conclusion Experimental results show that the proposed algorithm can effectively overcome the privacy leakage of facial image due to publication, the released facial images are accurate, and satisfy ε–differential privacy. The algorithm has good robustness, and it is an effective private facial image releasing method.