基于形状纹理的JPEG图像检索系统新策略

19 浏览量更新于2024-08-26 收藏 294KB PDF 举报

本文档探讨了"基于形状和纹理水印的彩色图像检索系统"，这是一项在多媒体技术领域内具有前瞻性和增长潜力的研究。内容基于内容的图像检索（CBIR）是该领域的核心课题，它通过分析图像内容而非元数据来实现精确的信息搜索。本文作者 Hao Zhang、Hua Chen、Fa-Xin Yu 和 Zhe-Ming Lu 来自浙江大学航空航天学院，他们的研究聚焦于如何将水印技术融入到JPEG图像检索系统中，提出了一种新颖的方法。在该系统的构建中，首先进行的是离线处理阶段，也称为预处理步骤。在这个阶段，每个图像被提取特征并转化为水印形式嵌入到原始图像中，而无需解压缩JPEG图像，而是直接在离散余弦变换（DCT）域中嵌入。这种方法简化了流程，提高了效率，因为特征计算是在水印嵌入过程中完成的。在线检索过程则包括三个关键步骤：查询特征计算、水印提取以及特征向量匹配。由于特征已经被嵌入到图像数据中，因此在检索时，只需要提取出嵌入的水印，并与查询图像中的水印进行比较，而不必重新计算特征。这种特性显著降低了检索时间，并提高了检索精度，因为它直接利用了图像数据中的内在信息。本文的创新之处在于结合了形状和纹理两个维度的特征，以增强图像的识别性和区分度，使得检索系统能够更好地识别和区分具有相似外观但不同形状或纹理特征的图像。这项研究对于提高图像检索的准确性和效率具有重要意义，为多媒体技术的发展提供了新的思路和解决方案。

Color Image Retrieval System Based on Shape and

Texture Watermarks

Hao Zhang, Hua Chen*, Fa-Xin Yu and Zhe-Ming Lu

School of Aeronautics and Astronautics

Zhejiang University

Hangzhou 310027, China

Email: chen2013571@126.co m; zheminglu@zju.edu.cn

Abstract—Content-based image retrieval (CBIR) is one of the

most exciting and fastest-growing areas in the field of multimedia

technology. In this Letter, we apply the watermarking technique

into the retrieval system and propose a novel approach for JEPG

image retrieval. The proposed image retrieval system consists of

two main phases, offline process and online retrieval process. The

feature vector is extracted from each image as the watermark to

be embedded into the image, which is the preprocessing

operation called offline process. It doesn’t require decompressing

the JPEG images but directly embedding the watermark in the

DCT domain. The online retrieval process consists of three

processes, i.e., query feature computation, watermark extraction

and feature vector matching. Since the features are embedded in

the image data, it is unnecessary to compute the features but only

to extract it from the watermarked image. We carry out a series

of experiments on a watermarked image database, and the

detailed comparison results between the novel approach and the

traditional system indicate the advantage of the proposed method.

Keywords-image retrieval; content-based image retrieval;

digital image watermarking

I. INTRODUCT ION

With the explosive growth of the image databases in terms

of both the size and the variety, more and more attention has

been paid to the research of automatic image retrieval. Content-

based image retrieval (CBIR), which was proposed in the early

1990’s[1], has been one of the most active research areas in the

past two decades. Many image features, such as color, texture

and shape have been explored to describe the content of the

images and many CBIR systems have been established. Most

existing CBIR systems focus primarily on feature analysis,

similarity measures and feedback learning algorithms [2], as

shown in Fig.1. However, there is no significant effort to

investigate the security problem in CBIR, except Forsyth et

al.’s filter for detecting naked people [3]. Actually, with the

rapid development of digital metadata over the Internet, the

security problem is becoming more and more obvious. For

example, for a remote sensing image database or a face

database, the database has to be open to all the uses without

any authorization [4].

Digital watermarking technology, as an approach of

embedding some information into the image for solving the

security problem, was formed in the middle 1990’s. Digital

watermark is the infor mation ins erted into the host digital

multimedia such as image, video, audio and text. The process

to insert a digital watermark is called digital watermarking.

Digital watermarking is an important approach to ensure the

security of information. It may be widely used in copyright

protection, labeling, monitoring, tamper proofing, conditional

access, national defense, national security, and so on. It has

been a hotspot in the information security field.

Figure 1. T he traditional retrieval system.

Therefore, considering the security problem in the evolving

CBIR systems, watermarking techniques used in CBIR have

been investigated in recent years [4, 5]. In this Letter, we

combine the two relatively new research fields together and

present a novel retrieval system based on the watermarking

technique. The feature vectors are computed as watermarks and

invisibly embedded into images, by which the watermarked

image database is formed during the offline processing phase.

Traditionally, during the offline process, the image retrieval

system for JPEG images needs firstly decompressing the

images and then performing the feature extraction in the spatial

The project is supported by the National Natural Science Foundation of

China under grant No. 61171150 .

Visual cont

ent

Description

Feature

Vectors

Similarity

Match

Feature

Database

Visual

Content

Description

Image

Database

Indexing &

Retrieval

Results

Relevance

Feedback

Query

Formation

User

Output

2012 Second International Conference on Instrumentation & Measurement, Computer, Communication and Control

DOI 10.1109/IMCCC.2012.141

569

下载后可阅读完整内容，剩余3页未读，立即下载

weixin_38622849

粉丝: 3
资源: 958

基于形状纹理的JPEG图像检索系统新策略

基于颜色和纹理特征的图像检索

基于纹理特征的图像检索系统

基于颜色和纹理特征的图像检索研究

基于纹理特征的图像检索系统设计

基于形状和纹理的外观设计专利图像检索方法 (2011年)

基于颜色和纹理特征的图像检索技术研究 (2008年)

基于分块主色调和纹理特征的彩色图像检索.pdf

基于颜色和纹理的图像检索

基于颜色空间和纹理特征的图像检索

基于颜色特征、基于形状特征或者基于颜色和形状综合特征图像检索

最新资源