Development of an Enhanced Bit Plane Complexity Segmentation Based Homomorphic Encryption for Privacy Preservation in Distributed System

Privacy preservation in distributed systems faces persistent challenges in securing sensitive data while maintaining confidentiality and image quality. Conventional Bit Plane Complexity Segmentation (BPCS) techniques use fixed complexity thresholds, limiting embedding efficiency, and they do not protect embedded data if detected. This study developed an Enhanced BPCS based Homomorphic Encryption (EBPCS+HE) framework integrating an adaptive threshold-based region selection mechanism with the Paillier homomorphic encryption scheme to enhance data confidentiality and enable encrypted-domain operations. Implemented in MATLAB R2023a using thirty Chest X-ray images, the framework was evaluated using PSNR, payload ratio, encryption/decryption time, memory usage, and Shannon entropy. The developed framework achieved an average PSNR of 50.75 dB, payload ratio of 23.48%, embedding, encryption, and decryption times of 2.54 s, 1.80 s, and 1.41 s, respectively, and encryption memory usage of 2.92 MB. Comparative analysis showed improved performance over existing BPCS-based methods. Average Shannon entropy increased from 7.4168 in cover images to 7.8934 after encryption, indicating enhanced randomness and resistance to statistical attacks. The study concludes that the EBPCS+HE framework provides an effective solution for privacy preservation in distributed systems by improving embedding capacity, preserving image quality, strengthening data confidentiality, and enabling processing operations on encrypted data.