In this paper, a cloud based digital image locker system along with secure user authentication and novel image cryptosystem for preserving privacy of user images is proposed. Images require more storage space as compared to text. So, it is not feasible to store images locally in digital locker for everyone. By taking advantage of high speed Internet, it is feasible to use cloud for efficient storage of images. However, so many researchers feel that cloud technology is ‘honest-but-curious’. Cloud service follows the specified protocol completely, but it traces and analyzes the data to get the user’s important information. To overcome from such challenges, privacy of user’s identity and data should be preserved. The proposed authentication protocol is designed using Paillier based difference function which is based on Homomorphic cryptosystem. The speciality of homomorphic function is that computation on encrypted data is possible without decrypting the cipher text. Because of this, authentication process is done in encrypted environment. A novel image cryptosystem is designed based on Fridrich model. In this cryptosystem, confusion is performed at bit-level and diffusion is performed at pixel-level. Keys are modified in each round of confusion to achieve more security. Simulation and cryptanalysis prove the effectiveness of the proposed system.

在本文中，提出了一种基于云的数字图像储物柜系统以及安全的用户身份验证和新颖的图像密码系统，以保护用户图像的隐私。与文本相比，图像需要更多的存储空间。因此，将图像本地存储在每个人的数字储物柜中是不可行的。通过利用高速Internet，使用云来有效存储图像是可行的。但是，许多研究人员认为云技术是“诚实而又好奇的”。云服务完全遵循指定的协议，但是它会跟踪和分析数据以获取用户的重要信息。为了克服这些挑战，应保留用户身份和数据的隐私。提出的认证协议是使用基于Paillier的差分函数设计的，该差分函数基于同态密码系统。同态函数的特殊之处在于，无需解密密文就可以对加密数据进行计算。因此，认证过程是在加密环境中完成的。基于Fridrich模型设计了一种新颖的图像密码系统。在该密码系统中，混淆在比特级别执行，扩散在像素级别执行。在每一轮混乱中都会修改密钥，以提高安全性。仿真和密码分析证明了该系统的有效性。混淆是在位级别执行的，扩散是在像素级执行的。在每一轮混乱中都会修改密钥，以提高安全性。仿真和密码分析证明了该系统的有效性。混淆是在位级别执行的，扩散是在像素级执行的。在每一轮混乱中都会修改密钥，以提高安全性。仿真和密码分析证明了该系统的有效性。