Reversible Data Hiding Scheme Based on Quad-Tree and Pixel Value Ordering

This study proposes a reversible data hiding (RDH) scheme based on pixel value ordering (PVO). RDH schemes based on PVO technique generally realize the prediction procedure by employing the sorted pixels in a block-wise manner, and then embedding secret data by expanding the prediction errors. Since...

Full description

Saved in:
Bibliographic Details
Published inIEEE access Vol. 7; pp. 142947 - 142962
Main Authors Li, Jian-Jun, Lee, Chin-Feng, Chang, Chin-Chen, Lin, Jiang-Yi, Wu, Yun-He
Format Journal Article
LanguageEnglish
Published Piscataway IEEE 2019
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects
Complexity
Complexity theory
Correlation
Embedding
Histograms
Image quality
Image segmentation
Payloads
Pixel value ordering
Pixels
prediction error expansion
Prediction methods
quad-tree
reversible data hiding
Online AccessGet full text

Cover

More Information
Summary:This study proposes a reversible data hiding (RDH) scheme based on pixel value ordering (PVO). RDH schemes based on PVO technique generally realize the prediction procedure by employing the sorted pixels in a block-wise manner, and then embedding secret data by expanding the prediction errors. Since these operations are carried out based on the block, and pixel intensity has spatial correlation such that the neighboring pixels have similar values, the block patterns and sizes certainly affect the embedding performance of the PVO-based RDH method. The original PVO-based RDH schemes use fixed block size. However, due to the differences in image complexity, fixed-size blocks have many limitations. Thus far, several methods that break the limitations have been proposed, such as setting a threshold value, and determine the size from the complexity of the block, using sliding window instead of fixed blocks. These methods significantly enhanced the embedding performance of PVO-based RDH schemes, but this can still be improved. The proposed scheme introduces the quad-tree structure, which combines the advantages of several previously proposed methods to design a flexible block patterns, thus fully utilizing the characteristics of the image itself in embedding secret data. Experimental results indicate that the proposed scheme is flexible and better than previous approaches.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2019.2941500