本文研究了用于數據嵌入的預留空間。因為低壓縮率不能提供足夠的空間來控制額外的比特,而預留空間為數據嵌入提供額外的空間,所以預留空間在最大化數據嵌入中起關鍵作用。但小塊分區(qū)上的結果不能容納該塊所需的足夠數據,而在較低的塊級別上,可以獲得最高的嵌入容量,因此,基于動態(tài)分區(qū)的平滑區(qū)域和粗糙區(qū)域,對于以最大的嵌入率尋找到最佳的信息寫入位置和保持高質量的圖像具有重要作用。首先,本文提出了一種基于多比特平面高效壓縮域的可逆數據隱藏新技術。本文通過一系列實驗,利用分塊方案對不同參數的結果進行了評價,并修正了直方圖各分塊的零點概率。該方案獲得了更多的嵌入容量和高質量的隱寫圖像。實驗結果有效地實現了高嵌入能力和保持圖像質量的目標。其次,本文提出了一種創(chuàng)新的可逆數據隱藏技術,該技術利用多層局部化n位截斷圖像(LBPTI)在柱狀圖移動的基礎上形成的,即通過有效無損壓縮從8位平面生成。從塊中選擇參考點后,利用相鄰的頂點,在不修改峰值點的情況下實現數據嵌入;另外,在提取端,提取秘文信息時,對峰值的關鍵信息并不是強制性提取的。為了使嵌入的覆蓋圖像與原始覆蓋圖像的直方圖相似,在保持覆蓋圖像質量外,同時本文還利用低塊...

【文章頁數】：110 頁

【學位級別】：博士

【文章目錄】：
Acknowledgements
摘要
Abstract
1 Introduction
    1.1 Background and Scope
    1.2 Motivation
    1.3 Research methodology
    1.4 Dissertation Organization
2 A new Multilevel Reversible Bit-planes Data Hiding Technique Based on His-togram Shifting of Efficient Compressed Domain
    2.1 Related Work
        2.1.1 Ni et al.'s method
        2.1.2 Kim et al.'s method
        2.1.3 Che et al.'s method
    2.2 Proposed method
        2.2.1 Data embedding algorithm
        2.2.2 Data extraction and cover image retrieval algorithm
    2.3 Experiments
        2.3.1 Performance comparisons
        2.3.2 Block Divisions
        2.3.3 Computational Complexity
        2.3.4 Lower bound PSNR
    2.4 Conclusions
3 Efficient Lossless Compression based Reversible Data Hiding using Multilay-ered n-bit Localization
    3.1 Related work
        3.1.1 Kim et al.'s Scheme
        3.1.2 Che et al.'s Scheme
    3.2 Data embedding algorithm
        3.2.1 Embedding with n-bit localization
        3.2.2 Side information
        3.2.3 Data extraction algorithm
        3.2.4 Multilayer n-bit embedding
    3.3 Experiments
        3.3.1 N-bit plane with different embedding layer
        3.3.2 Comparison with existing algorithms
        3.3.3 Embedded capacity versus PSNR
    3.4 Conclusions
4 Generalized PVO base Reversible Data Hiding Using Firefly Algorithm
    4.1 Introduction
    4.2 Related work
    4.3 Proposed Schemes
        4.3.1 Quadtree image partition
        4.3.2 Embedding scheme
        4.3.3 Extraction scheme
        4.3.4 Data embedding and extraction procedures
    4.4 Experiments
        4.4.1 Analysis of Proposed Method
        4.4.2 Special Blocks Handling
        4.4.3 Performances Comparison
    4.5 Conclusions
5 RDH based histogram equalization using for Cancer prediction and recognitionfor Abnormal Tumor regions
    5.1 Related work
        5.1.1 Histogram smoothness by Gaussian filter
        5.1.2 Partition of new dynamic range
        5.1.3 Independently equalized each partition
        5.1.4 Normalization of image brightness
    5.2 Limited Dynamic weighted histogram equalization
        5.2.1 Image Preprocessing
        5.2.2 Data embedding and extraction procedures
    5.3 Experiments
        5.3.1 Analysis of Proposed Method
        5.3.2 Performance analysis for disease classification
    5.4 Conclusions
6 Conclusions and Future Directions
References
Publications



本文編號：3837668