Enjoying the rapid growth of image data and cloud computing platforms, various image processing ap-plications have emerged and flourished in recent years. Meanwhile, the privacy concerns over the abuse of sensitive information contained in outsourced data also arise in public. In fact, once uploaded to the cloud, the security of us-ers’ private information purely depends on the reliability of the cloud service providers (CSP). To solve this problem, the security requirements and technical challenges lain in privacy-preserving image processing based on different cloud computing architectures were studied, and several solutions to protect the security of outsourced data while enabling functionality of image processing applications were proposed. Several state-of-the-art techniques for secure image processing were introduced and analyzed, including homomorphic encryption (HE) scheme, secure multiparty computation (SMC) protocol, and differential privacy (DP).