Medical imaging is crucial for identifying diseases, and strict security and privacy regulations need to be implemented because of the sensitive nature of these pictures. For Healthcare Industry 4.0, medical pictures in cloud-based medical systems should be safeguarded before being outsourced. Nevertheless, processing queries over encrypted data without first completing the decryption stage is presently both challenging and impracticable. In the study, we provide a practical way to find the exact closest neighbour among a collection of encrypted medical images. Rather of calculating the Euclidean distance, we may reject candidates by determining the lower limit of the distance, which is connected with the data's mean and standard deviation. Unlike most other current algorithms, our method is able to locate the exact closest neighbour instead of an approximation. Next, we evaluate our proposed approach to demonstrate its efficacy