加密JPEG位流中的隐私保护信号处理与可逆数据隐藏

65 浏览量更新于2024-08-28 收藏 1.39MB PDF 举报

本文主要探讨了隐藏在加密JPEG位流中的可分离可逆数据的研究，针对近年来日益增长的信号处理在加密领域（SPED）隐私保护需求。随着云计算和远程计算的发展，用户倾向于将原始信号的加密副本发送到服务器，以保护数据隐私，而服务器则需在加密状态下完成信号处理任务。作者们提出了一系列创新方法，旨在应对不同的应用场景，如加密图像压缩、基于密文的信号变换、加密环境下的模式识别以及加密多媒体水印和数据搜索等。首先，文章的焦点在于介绍一种新的技术，即如何在加密的JPEG图像数据流中嵌入可分离可逆的信息，这使得数据能在保持其保密性的同时，仍然可以进行必要的处理和分析。JPEG是一种广泛应用于数字图像压缩的标准，通过量化和熵编码将连续的像素数据转化为离散的位流。在传统的JPEG图像处理中，这些步骤通常是不可逆的，但在加密环境下，研究人员找到了一种方式来实现信息隐藏，同时保持了数据的可恢复性。 1. 加密图像压缩：研究者可能设计了一种算法，允许在压缩过程中悄悄地嵌入数据，而在解压缩时，这些隐藏的信息能够完整恢复。这种方法有助于保护图像的原始内容，同时支持服务器进行诸如图像识别或内容分析等操作。 2. 密文信号处理：在文本或其他信号中，加密后的数据可能通过特定的数学变换进行处理，比如傅立叶变换或小波变换。在这些变换中，研究人员可能发现了利用密文进行信号分析的策略，同时确保处理结果对原始信息的保密性影响最小。 3. 加密模式识别：在加密环境下进行模式识别，例如人脸识别或语音识别，需要处理的不仅仅是加密的像素值，还有可能涉及到数据的秘密特征。研究者可能会开发出一种能够在加密模式下提取关键特征的方法，以便于在保护用户隐私的前提下，服务器仍能识别出模式。 4. 加密多媒体水印：为了防止未经授权的复制或篡改，文章可能讨论了如何在加密的音频或视频中嵌入难以检测但可逆的水印，这在版权管理和防伪上具有重要意义。 5. 数据搜索在加密环境中：考虑到数据安全性和效率，研究者可能探索了如何在不暴露实际数据的情况下，在大型加密数据库中执行高效的数据查询，这涉及到加密数据的索引构建和搜索算法优化。这篇2016年的IEEE Transactions on Dependable and Secure Computing的研究论文提供了一种创新的思路，将信号处理与加密技术相结合，以满足隐私保护的需求，同时也推动了信息安全领域的理论和技术发展。

1545-5971 (c) 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See 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/TDSC.2016.2634161, IEEE

Transactions on Dependable and Secure Computing

embedding rates and good recovery capability. However, they

all require extra RDH operations before image encryption,

thus contradict the very purpose of RDH-EI, since the extra

operations are performed to the plaintext rather than encrypted

images.

There are also some interesting works based on

commutative encryption and data hiding [24-26], where

commutative means that the orders of encryption and data

hiding/extraction can be swapped. Although some have

separable features, commutative RDH is different from the

framework proposed in the present paper.

B. Joint RDH-EI

In joint RDH-EI, the additional message can only be

extracted by the recipient after image decryption, along with

image recovery, while the data hider cannot perform

extraction.

A feasible method was first proposed in [8], in which the

content owner encrypts an original image using a stream

cipher, and the data hider embeds additional messages into

ciphertext blocks by flipping three least significant bits (LSB)

of half the pixels in each block. When extracting the additional

messages, the recipient decrypts the marked encrypted image

and generates two candidates for each block by flipping LSBs

again. Since the original block is much smoother than the

interfered, the embedded bits can be extracted and the original

image perfectly recovered. This joint RDH-EI method

depends on the size of each block. As long as the block size is

appropriately chosen, errors of extraction and recovery can be

avoided. This method was improved in [9] by exploiting

spatial correlation between neighboring blocks and using a

side-match algorithm to achieve a higher embedding rate. The

flipping based approach was further improved in [10], in

which multiple neighboring pixels in different locations are

used to reduce error rates in extraction and recovery.

Recently, a new joint RDH-EI method was proposed in [11].

Data embedding is realized through a public key modulation

mechanism. On the recipient end, a two-class SVM classifier

is designed to distinguish encrypted and non-encrypted image

patches. Consequently, the recipient can jointly extract the

additional messages and recover the original image. This

method provides a higher embedding capacity.

C. RDH-EI for JPEG Bitstream

As most RDH-EI methods are designed for uncompressed

spatial-domain images, [12] proposes an approach capable of

reversely hiding messages into encrypted JPEG bitstreams.

This scheme aims at encrypting a JPEG bitstream into a

properly organized structure and embedding additional

messages into the encrypted bitstream by slight modifications.

During the bitstream encryption, all appended bits of the

Huffman codes are encrypted with a stream cipher, and all

Huffman codes are kept unchanged. After encryption, the file

size is preserved, and the format is compliant to common

JPEG decoders. On the server side, the bitstream of every

other block is selected as a candidate. If all AC coefficients of

a candidate block are zero, the block is skipped. Additional

bits are then encoded by LDPC-based error correction codes

(ECC), and embedded into the useful candidate bitstream by

flipping the LSBs of the encrypted appended bits of the AC

coefficients in each candidate block. On the recipient side,

LSBs of the appended bits of each candidate bitstream are

flipped again to estimate the additional bits using a predefined

blocking artifact function and an ECC decoder. Meanwhile,

the original bitstream can be losslessly recovered according to

the extracted bits.

In [27] and [28], some interesting ideas of RDH were

proposed for JPEG images by combining image scrambling

and data embedding. By scrambling the JPEG structure,

additional message is embedded into the encrypted bitstream.

However, in these methods data embedding must be combined

with image encryption, which is different from general

RDH-EI framework depicted in Fig. 2.

Limited by JPEG compression, large embedding capacity

cannot be achieved. In [12], about 750 bits are embedded into

the JPEG bitstream of a 512×512 grayscale image. The joint

RDH-EI method requires a combined data extraction and

image recovery. That may become a problem since the

database administrator cannot read the hidden messages from

the marked encrypted bitstream. As format compliance is

required in JPEG encryption [29], it is difficult to design a

secure encryption algorithm for JPEG. The algorithm

previously presented in [12] is not secure enough. Analyses in

[27] show that the principal structure of the original image can

be estimated from the encrypted bitstream if all Huffman

codes are kept unchanged. In view of these drawbacks, we

provide a new encryption scheme for JPEG bitstream, and

propose a separable RDH-EI approach for the encrypted

bitstream. In the proposed method, data extraction and image

recovery are separated, higher embedding capacity is achieved,

and security of JPEG encryption enhanced.

III. PROPOSED FRAMEWORK

The framework of the proposed method is depicted in Fig. 3.

The JPEG RDH-EI workflow includes three parties: content

owner, data hider, and recipient.

Given a JPEG bitstream and an encryption key, the content

owner generates a ciphertext bitstream after syntax parsing

and encryption. In the process, the file size is kept unchanged

and the format is compliant to common JPEG decoders.

When a remote server receives the encrypted bitstream, the

data hider parses the bitstream and hides additional messages

in it using an embedding key. After the marked encrypted

bitstream is constructed, the file size and format compliance

are preserved. In this scheme, the server can extract additional

messages from the marked encrypted bitstream using the

embedding key.

On the recipient side, the additional messages can also be

extracted from the received bitstream if the embedding key is

available. A recipient with only the encryption key can view

an approximate image by a direct decryption. If both the

encryption and embedding keys are available, the recipient can

losslessly recover the original bitstream after decrypting the

marked encrypted JPEG bitstream.

剩余12页未读，继续阅读

weixin_38694800

粉丝: 4
资源: 1021

加密JPEG位流中的隐私保护信号处理与可逆数据隐藏

自适应编码在加密图像大容量可逆数据隐藏中的应用

MATLAB实现：加密图像中的可逆数据隐藏技术

改进的可分离密文域可逆信息隐藏：提升安全与可逆性

winhex数据恢复小脚本

【Commons-Codec高级应用】：定制化数据处理流程，打造高效工作流

Entity Framework在微服务架构中的角色：分布式数据处理的最佳实践

组态王系统效率优化：数据压缩与传输的绝招

【离散余弦变换（DCT）】：数字信号处理中不可或缺的关键技术

【Python第三方库应用】：探索base64在扩展库中的应用与集成技巧

加密医学图像的可逆信息隐藏算法研究

最新资源