JPEG图像中的可逆数据隐藏技术：挑战与应用

需积分: 9 98 浏览量更新于2024-08-26 收藏 1.83MB PDF 举报

本文主要探讨了隐藏在JPEG图像中的可逆数据这一主题，JPEG作为日常生活中最常见的数字图像格式之一，其广泛应用在诸如档案管理、图像认证等场景中，因此可逆数据隐藏（Reversible Data Hiding, RDH）技术在JPEG图像中的研究具有重要意义。然而，与未压缩图像相比，JPEG图像的处理更具挑战性。由于JPEG压缩过程中信息冗余较低，任何在压缩域内的修改都可能导致宿主图像的更多失真。这就需要设计高效的算法，能够在保持图像质量的同时实现数据的有效隐藏。文章的核心关注点在于如何在JPEG编码的图像中实现高效率和低失真的数据嵌入。这涉及到以下几个关键知识点： 1. **JPEG压缩特性**：JPEG利用有损压缩技术，通过DCT（离散余弦变换）和量化过程来减小图像数据量，牺牲部分图像细节以换取压缩比。这种特性使得在已压缩的图像中嵌入数据时，需要考虑如何在不影响视觉质量的前提下，合理利用JPEG的内在结构。 2. **可逆性与不可逆性**：在JPEG图像中进行RDH，必须确保数据隐藏是可逆的，即数据隐藏后可以通过某种方式恢复原图像，而不会造成明显的失真。这需要嵌入算法具备一定的逆向操作能力，如选择合适的嵌入位置、调整量化参数或利用边缘、纹理等视觉冗余。 3. **容量与质量权衡**：为了在有限的容量内隐藏更多的数据，算法需要在嵌入数据量和图像视觉质量之间寻找最佳平衡。这通常涉及复杂的编码策略和解码算法，以最小化嵌入对原始图像的感知影响。 4. **安全性与鲁棒性**：除了数据隐藏，图像认证也是应用中的一个关键需求。文章可能探讨了如何在保证数据安全性和隐藏信息难以被检测到的同时，使图像的认证过程更为可靠。 5. **标准与研究进展**：提到的“DOI 10.1109/TCSVT.2015.2473235”表明这是一篇发表于《IEEE Transactions on Circuits and Systems for Video Technology》的研究论文，可能是该领域的最新研究成果，介绍了当时关于JPEG图像可逆数据隐藏的最新技术成果和未来研究方向。这篇文章深入研究了在JPEG图像中实施可逆数据隐藏的复杂性，以及如何通过创新算法和技术解决这一难题，以满足实际应用的需求。

http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI

10.1109/TCSVT.2015.2473235, IEEE Transactions on Circuits and Systems for Video Technology

T-CSVT-9331

two-dimensional DCT to the non-overlapping 8×8 blocks, the

original image data are transformed from the spatial domain to

the frequency domain. The obtained DCT coefficients are then

fed to the quantizer and quantized using the predetermined

quantization table. The quantized DCT coefficients are

arranged in a zigzag scanning order and pre-compressed using

the differential pulse code modulation (DPCM) on direct

current (DC) coefficients and run length encoding (RLE) on

AC coefficients. Finally, the symbol string is Huffman coded

to obtain the final compressed bitstream. After pre-pending the

header, we obtain the final JPEG file.

In the proposed scheme, we only choose to embed message

into AC coefficients (See some explanations to be provided

next). Thus, in the following paragraphs, DPCM on DC

coefficients is neglected and only RLE for AC coefficients is

described in detail.

Fig. 1. Block diagram of JPEG compression process.

The AC coefficients are encoded in a specific RLE format

as intermediate symbols (R/C,V). Here, R ( 0 ≤R ≤15 )

denotes the zero run length (i.e., the number of zero AC

coefficients before the next nonzero AC coefficient), and C

(1 ≤C≤10) denotes the category of the next nonzero AC

coefficients (i.e., the number of bits needed to represent the

amplitude of the next nonzero AC coefficient), which is shown

in Table I. There are 160 combinations of R and C. Each R/C is

encoded with a variable-length code (VLC) from the Huffman

table; the code length ranges from 1 to 16. In addition, because

the run length of zero coefficients may exceed 15, the value

R/C F0H is defined to represent a run length of 15 zero

coefficients followed by a coefficient of zero amplitude (this

can be interpreted as a run length of 16 zero coefficients), and

a special value R/C 00H is used to code the end-of-block

(EOB) when all remaining coefficients in the block are zero,

where “H” implies that “F0” and “00” are hexadecimal

numbers. Thus, there are in total 162 different VLC codes in

the standard Huffman table.

TABLE I

CATEGORIES ASSIGNED TO COEFFICIENT VALUES

AC Coefficients

-1,1

-3,-2,2,3

–7...–4,4...7

–15...–8,8...15

–31...–16,16...31

–63...–32,32...63

–127...–64,64...127

–255...–128,128...255

–511...–256,256...511

–1023...–512,512...1023

Note that all R/C values are encoded with variable-length

Huffman codes, but the V values are not Huffman coded. Each

V value is encoded with a variable-length integer (VLI) code,

whose length in bits is given in the first column of Table I.

They are appended to the Huffman-coded R/C to form the final

JPEG bitstream. For more detail, please refer to the JPEG

guidelines released by International Telecommunication

Union [30].

B. Selection of Coefficients for RDH

Almost all state-of-the-art HS-based RDH algorithms

consist of three steps. The first step is to generate a sequence

with small entropy from the host image, which can be realized

by using the original histogram of the host image directly, the

difference histogram, or the prediction error histogram. The

second step is to divide the obtained histogram into two

regions. One is called the inner region, and the other is called

the outer region. In general, some bins associated with the

peak points are divided into the inner region and the rest of the

bins are divided into the outer region. The third step is to

reversibly embed the message. That is, the bins in the inner

region are expanded to carry data, whereas these in the outer

region are shifted so that the inner and outer regions are still

separate from each other after data hiding. Note that the

expansion and shifting of bins are accomplished by modifying

the corresponding pixel/coefficient values of the host image.

For ease of presentation, in the following discussion, the

pixels/coefficients corresponding to the elements in the inner

region and the outer region are called inner pixels/coefficients

and outer pixels/coefficients, and the modifications made on

the inner and outer pixels/coefficients are called expansion and

shifting, respectively.

For JPEG images, the quantized DCT coefficient histogram

can be easily obtained, and the message can be embedded into

the JPEG images with a (DCT coefficient) histogram shifting

strategy directly. However, because of the de-quantization

process, any modification of the quantized DCT coefficients

may introduce considerable distortion in the spatial domain,

and it may also significantly increase the JPEG file size. In the

following paragraphs, we will present an analysis on which

types of coefficients can be selected for data hiding. For the

ease of explanation, we resort to the standard 512×512 Lena

image with different quality factors in our discussion, which is

the most popularly utilized image for testing the efficiency of

different RDH schemes currently.

1) About DC coefficients

As pointed out by Reininger and Gibson [31], for most of

the JPEG images, the statistics of the DC coefficients are best

approximated by a Gaussian distribution.

The histograms of all quantized DC coefficients of the Lena

image with different quality factors (i.e., QF = 70, 80, 90, 100)

are shown in Fig. 2. As shown, the histograms of DC

Original

Image Data

(8×8 blocks)

Compressed

Image Data

JPEG Encoder

Quantizer

DCT

Entropy

Encoder

剩余11页未读，继续阅读

weixin_38599545

粉丝: 7
资源: 935

JPEG图像中的可逆数据隐藏技术：挑战与应用

基于DCT频率和块选择的JPEG图像中可逆数据隐藏

电信设备-基于二维直方图修改的JPEG图像可逆信息隐藏方法.zip

JPEG压缩的可逆信息隐藏

JPEG图像压缩原理

MATLAB显示jpeg图像

能用C语言写一段将YUV图像转换为JPEG图像的代码吗

基于matlab的jpeg图像压缩编解码

Jpeg 图像压缩原理

jsteg-jpeg

jpeg图像压缩代码matlab

最新资源