REVERSIBLE DATA HIDING IN ENCRYPTED DOMAIN BY USING HISTOGRAM SHIFTING AND COMPRESSION TECHNIQUES

This paper presents a novel reversible data hiding approach in the encrypted domain (RDHED) that combines block-adaptive histogram shifting with context-aware compression to substantially enhance embedding capacity while guaranteeing complete reversibility of both the image and the embedded data. In contrast to conventional global histogram shifting methods, the proposed technique partitions the encrypted image into entropy-driven adaptive blocks, allowing localized peak–zero bin selection for efficient data embedding. To further increase available embedding space, each block is subjected to a dual-mode lossless compression strategy, dynamically selecting between Run-Length Encoding and Huffman coding based on local statistics. The encryption stage, implemented using AES in counter mode (AES-CTR), preserves the intensity distribution, enabling effective histogram-based embedding in the encrypted domain. This paper also provides mathematical formulations for embedding capacity, distortion analysis, and compression gain, as well as computational complexity evaluation. Experimental results on standard benchmark images demonstrate that the proposed approach achieves up to 28% higher embedding capacity compared to existing RDH-ED schemes, with a PSNR degradation of less than 1 dB and maintaining 100% recovery accuracy. The method is fully implemented in Python, ensuring reproducibility and ease of extension to colored images and video content. This paper shows that the proposed algorithm offers a practical and secure solution for applications such as privacy-preserving cloud storage, secure medical image annotation, and digital forensics.