CAJ | 학술논문

现有基于奇异值分解(SVD)的彩色信息加密系统提供了一种光学矩阵分解方案、安全的密文和敏感的密钥。高维张量奇异值分解(HOSVD)是SVD矩阵的自然线性延伸，提出了一种基于HOSVD的彩色图像加密算法。在加密过程中，HOSVD比SVD提供了更多的密文乘法组合次序。这些乘法组合次序可以有效地增加未经授权的解密难度。在解密过程中，HOSVD的重建精度比SVD更高。这些优点提高了准确性、安全性和鲁棒性。通过对100个图像测试数据集的计算机仿真验证了该算法的可行性。
현유기우기이치분해(SVD)적채색신식가밀계통제공료일충광학구진분해방안、안전적밀문화민감적밀약。고유장량기이치분해(HOSVD)시SVD구진적자연선성연신，제출료일충기우HOSVD적채색도상가밀산법。재가밀과정중，HOSVD비SVD제공료경다적밀문승법조합차서。저사승법조합차서가이유효지증가미경수권적해밀난도。재해밀과정중，HOSVD적중건정도비SVD경고。저사우점제고료준학성、안전성화로봉성。통과대100개도상측시수거집적계산궤방진험증료해산법적가행성。
The existing Singular Value Decomposition (SVD) based color information encryption system provided an optical matrix composition scheme, secure ciphertexts and very sensitive keys. As the Higher-order SVD (HOSVD) is a natural multi-linear extension of the matrix SVD, an HOSVD based color image encryption algorithm was proposed. In the encryption procedure, HOSVD can generate more multiplication orders of the ciphertexts (decomposition parts) than what SVD provides. These multiplication orders can be used as effective keys to make unauthorized decryption harder. In the decryption procedure, the reconstruction accuracy of HOSVD is higher than that of SVD. These advantages enhance the accuracy, security and robustness. Numerical simulations based on a test dataset of 100 images support the viability of the proposed algorithm.