基于词汇树层次语义模型的图像检索算法

解决语义鸿沟必须建立图像低层特征到高层语义的映射,针对此问题,本文提出了一种基于词汇树层次语义模型的图像检索方法.首先提取图像包含颜色信息的SIFT特征来构造图像库的特征词汇树,生成描述图像视觉信息的视觉词汇.并在此基础上利用Bayesian决策理论实现视觉词汇到语义主题信息的映射,进而构造了一个层次语义模型,并在此模型基础上完成了基于内容的语义图像检索算法.通过检索过程中用户的相关反馈,不仅可以加入正反馈图像扩展图像查询库,同时能够修正高层语义映射.实验结果表明,基于该模型的图像检索算法性能稳定,并且随着反馈次数的增加,检索效果明显提升.     

基于目标语义特征的图像检索系统

为克服当前基于内容的图像检索技术中低级特征无法准确全面地描述高级语义的问题,该文设计和实现了一个基于目标高级语义特征的检索系统。该系统利用了一个多级图像描述模型将语义特征结合到图像检索技术中。该图像描述模型通过在不同层次上对图像内容进行分析和描述,实现了从低级特征到高级语义的过渡。在此模型的基础上还研究了相应的检索机制和反馈技术。该系统的检索机制定位于图像中目标的语义内容,与传统的图像检索系统相比更接近人对图像内容的理解,从而使检索过程更简便,检索效率也得到很大提高。基于目标描述的自适应相关反馈可针对不同用户的不同需求给出相应的检索方案,从而使检索结果得到优化。     

基于概念格的图像语义检索研究

概念格是一种有效的数据分析和知识提取的形式化工具,已广泛应用于机器学习、人工智能、软件工程、知识发现等领域.提出了一种新的基于概念格的图像语义检索方法,将概念格理论应用到图像检索中,利用形式概念分析发现图像中潜在的概念结构和概念间的相互关系.借助于语言变量描述图像语义特征并根据这些模糊语义值构建概念格,用基于概念格的方法进行图像语义检索,这种方法所给的结果与人类视知觉具有更好的一致性.     

基于多级描述模型的渐进式图像内容理解

针对目前基于内容的图像检索技术中低级特征无法准确全面地描述高级语义的问题,本文提出了一种基于多级图像描述模型的渐进式图像内容理解.该图像描述模型在不同层次上对图像内容进行分析和提取,实现了图像内容的全方位描述,从底层向高层的过渡是渐进式的图像理解过程.特别是从视觉感知层到目标层,体现了图像低级特征与高级语义之间的过渡.本文给出了一种基于先验知识的上下文驱动的目标理解算法,实现了图像语义的提取.作为一个应用实例,本文给出了以上方法在基于内容的图像检索技术中的具体应用.     

基于中文自然语言的纹理分类新方法

提出了一种新的基于中文自然语言纹理描述词的纹理分类方法,建立了自然纹理分类体系,并用最小二乘支持向量机对纹理进行分类,实现了纹理的视觉特征到语义描述的转换.实验结果证明,该方法在图像理解和基于内容的图像检索中有助于缩小纹理特征的数学描述和人类理解之间的"语义鸿沟".     

基于人工神经网络的图像检索平台的设计与开发

用户描述图像的高层抽象语义与图像内在的底层特征之间存在差异,此时仅依靠图像内容特征进行检索的系统无法准确完成用户的检索任务。针对以上问题,提出了使用神经网络进行图像的匹配计算方法,通过样例自动学习和用户反馈学习两种学习方式,形成图像底层特征到图像分类的正确映射,学习后的神经网络可以进行图像的自动分类及检索。该方法结合了图像的底层特征描述及用户的高层语义反馈,有效地弥补了语义鸿沟。最后,系统通过整合Web前端、图像提取模块、神经网络模块及数据库模块,实现了神经网络学习及图像检索的完整流程。     

基于散布矩阵分析的相关反馈算法及应用

相关反馈技术是一种较常用的提高信息检索精度的方法.在图像检索领域,相关反馈技术被认为是解决图像高层语义内容和低层视觉特征之间差异的一种有效方法.视觉特征的权值调整是一类应用较多的相关反馈技术,权值调整方法中存在矩阵奇异问题,本文提出了一种新的基于散布矩阵分析的相关反馈算法,解决了矩阵奇异问题.该方法通过分析与检索目标相关图像在特征空间中的散布来构造目标图像类的投影空间,该空间对应于一个高层语义类在特征空间中分布密集的子空间,在投影空间中计算相似图像;同时根据每次反馈的信息不断修正投影空间来提高系统的检索性能.在Cord图像数据库中的实验结果表明该算法具有良好的检索性能.     

基于融合特征的现勘图像检索结果优化算法

刑侦现勘图像数据库是具有保密性高、图像内容罕见等极具行业特色的图像数据库.针对现勘图像内容复杂、目标物体不明确的特点,提出了DCT-DCT波纹理特征,并与HSV颜色直方图特征、GIST特征相融合构成融合特征.与常用的图像特征相比,DCT-DCT波纹理特征能够得到较高的检索效率,而融合特征的平均检索查准率高于构成其本身的三种特征的平均检索查准率.最后,将语义分析技术引入到检索过程中,提出基于检索结果优化的现勘图像检索算法,利用支持向量机（Support Vector Machine,SVM）分类器对查询图像进行语义提取,并对初次检索的结果进行语义分析,根据初检结果中语义类别的占比选择二次检索方案,该算法能在按例查询的基础上进一步提高平均检索查准率.     

基于“点到面”策略的图像检索

图像检索是计算机视觉领域的一个重要分支。其主要目的是从图像数据库中找出与查询图像相似的语义图像。传统的图像检索方法是在查询图像和数据库图像之间进行“点到点”检索。但是,单个查询图像包含的类别提示较少,即类别信息较弱,使得检索结果并不理想。为了解决这个问题,本文提出了一种基于“点到面”的类别检索策略来扩展一个图像（点）到一个图像类别（面）,这意味着从单个查询图像到整个图像类别的语义扩展。该方法挖掘了查询图像的类别信息。在两个常用的数据集上对所提出方法的性能进行了评估。实验表明,该方法可以显著提高图像检索的性能。      

基于视觉感知的图像检索的研究

基于内容图像检索的一个突出问题是图像低层特征与高层语义之间存在的巨大鸿沟.针对相关反馈和感兴趣区检测在弥补语义鸿沟时存在主观性强、耗时的缺点,提出了视觉信息是一种客观反映图像高层语义的新特征,基于视觉信息进行图像检索可以有效减小语义鸿沟;并在总结视觉感知的研究进展和实现方法的基础上,给出了基于视觉感知的图像检索在感兴趣区检测、图像分割、相关反馈和个性化检索四个方面的研究思路.     

JPEG图像重压缩检测综述

JPEG（Joint Photographic Experts Group,联合图像专家小组）是目前互联网上运用最为广泛的图像格式。已有的案例表明,许多篡改操作都发生在JPEG图像上,其操作基本流程是首先对JPEG文件进行解压,在空域进行篡改,篡改完成后再将篡改后的图片压缩保存为JPEG格式,这样篡改后的图片就可能会被两次甚至多次JPEG压缩。因此,JPEG图像的重压缩检测可以作为判断图像是否经过篡改的重要依据,对JPEG图像进行分析和取证具有非常重要的意义。本文主要从JPEG重压缩过程中量化表保持不变和量化表不一致这两个方面,对近年来JPEG重压缩检测领域的文献进行了一个回顾,介绍了该领域一些代表性的方法。最后我们还分析了JPEG重压缩领域存在的问题,并对未来的发展方向进行了展望。      

TMS320C54x Simulator文件格式及其应用

详细介绍了TMS320C54x Simulator环境的文件格式及其文件加载步骤,以及如何将BMP或JPEG格式灰度图像转换为能够使TMS320C54x Simulator识别的格式文件.实例应用表明此方法能够将图像信息加载到TMS320C54x Simulator环境中,进行相应的图像处理.     

File format technology in JPEG 2000 enables flexible use of still and motion sequences

While there exist many different image file formats, the JPEG committee concluded that none of those formats addressed a majority of the needs of tomorrow's complicated imaging applications. Many formats do not provide sufficient flexibility for the intelligent storage and maintenance of metadata. Others are very restrictive in terms of colorspace specification. Others provide flexibility, but with a very high cost because of complexity. The JPEG 2000 file format addresses these concerns by combining a simple binary container with flexible metadata architecture and a useful yet simple mechanism for encoding the colorspace of an image. This paper describes the binary format, metadata architecture, and colorspace encoding architecture of the JPEG 2000 file format. It also shows how this format can be used as the basis for more advanced applications, such as the upcoming motion JPEG 2000 standard.     

JPEG encryption for image rescaling in the encrypted domain

This work proposes a novel protocol of encrypting the JPEG image suitable for image rescaling in the encrypted domain. To protect the privacy of original content, the image owner perturbs the texture and randomizes the structure of the JPEG image by enciphering the quantized Discrete Cosine Transform (DCT) coefficients. After receiving the encrypted JPEG image, the service provider generates a rescaled JPEG image by down-sampling the encrypted DCT coefficients. On the recipient side, the encrypted JPEG image rescaled by the service provider can be decrypted to a plaintext image with a lower resolution with the aid of encryption keys. Experimental results show that the proposed method has a good capability of rescaling the privacy-protected JPEG file.     

基于JPEG图像文件格式的研究

针对最常见的JPEG图像格式的压缩数据进行分析,介绍了JPEG图像格式中的大部分典型且必要的标记;接着详细给出JPEG数据RGB与YUV之间转化的算法;最后,阐述了JPEG图像数据块内的编码方法。     

HD Photo的码率控制算法研究

HD Photo是微软公司开发的一种新颖的静止图像压缩算法和文件格式,并已通过JPEG组织提案,代号为JPEG XR,有望成为新的国际图像标准,其压缩性能堪比JPEG2000,计算复杂度却要低得多.然而,在目前公开的支持HD Photo的软件中,尚无一提供精确码率控制的功能.鉴于此,根据HD Photo格式的特性,提出了一种有效的基于二次R-D模型和线性回归预测的码率控制算法.实验证明,该算法能精确地控制图像的压缩大小,并且使图像质量接近于现有的DPK软件.     

Structural image retrieval using automatic image annotation and region based inverted file

Image retrieval has lagged far behind text retrieval despite more than two decades of intensive research effort. Most of the research on image retrieval in the last two decades are on content based image retrieval or image retrieval based on low level features. Recent research in this area focuses on semantic image retrieval using automatic image annotation. Most semantic image retrieval techniques in literature, however, treat an image as a bag of features/words while ignore the structural or spatial information in the image. In this paper, we propose a structural image retrieval method based on automatic image annotation and region based inverted file. In the proposed system, regions in an image are treated the same way as keywords in a structural text document, semantic concepts are learnt from image data to label image regions as keywords and weight is assigned to each keyword according to spatial position and relationship. As the result, images are indexed and retrieved in the same way as structural document retrieval. Specifically, images are broken down to regions which are represented using colour, texture and shape features. Region features are then quantized to create visual dictionaries which are similar to monolingual dictionaries like English or Chinese dictionaries. In the next step, a semantic dictionary similar to a bilingual dictionary like the English–Chinese dictionary is learnt to mapping image regions to semantic concepts. Finally, images are then indexed and retrieved using a novel region based inverted file data structure. Results show the proposed method has significant advantage over the widely used Bayesian annotation models.     

Content browsing and semantic context viewing through JPEG 2000-based scalable video summary

The paper presents a novel method and software platform for remote and interactive browsing of a summary of long video sequences as well as revealing the semantic links between shots and scenes in their temporal context. The solution is based on interactive navigation in a scalable mega image resulting from a JPEG 2000 coded key-frame-based video summary. Each key-frame could represent an automatically detected shot, event or scene, which is then properly annotated using some semi-automatic tools or learning methods. The presented system is compliant with the new JPEG 2000 Part 9 'JPIP - JPEG 2000 interactivity, API and protocols', which lends itself to working under varying transmission channel conditions such as GPRS or 3G wireless networks. While keeping the advantages of a single 2D video summary, like the limited storage cost, the flexibility offered by JPEG 2000 allows the application to highlight interactively key-frames corresponding to the desired content first within a low-quality and low-resolution version of the full video summary. It then offers fine grain scalability for a user to navigate and zoom into particular scenes or events represented by the key-frames. This possibility of visualising key-frames of interest and playing back the corresponding video shots within the context of the whole sequence (e.g. an episode of a media file) enables the user to understand the temporal relations between semantically related events/actions/physical settings, providing a new way to present and search for contents in video sequences.     

A requantization algorithm for the transcoding of JPEG images

Photos from digital camera are mostly lossy compressed in JPEG format, and almost all users do not save or cannot access the original image. Sometimes we need to reduce the size of this JPEG file for saving the disk space or for sending the image to the device with limited bandwidth and/or display size. Since the original image is not available, we need to recompress the already compressed photos. In the existing literature, it is revealed that the direct transcoding (decompressing the JPEG image and then compressing it again with a larger step size) generally results in lower quality images when compared to the one compressed from the original image with the same parameters. This discrepancy is due to the requantization error, and the existing algorithms for more efficient transcoding were focused on finding the requantization step sizes for reducing this error. In this paper, we propose an algorithm that manipulates the requantized DCT coefficients for enhancing the quality of transcoded image, regardless of the choice of requantization step sizes. For this purpose, we first locate the DCT coefficients which have the possibility of causing large requantization errors. But, since the exact requantization error cannot be computed without the original image, we define the error as the discontinuity of pixels around the block boundaries, between the given JPEG image and the targeting one. Then, for the possible variants of these coefficients, the errors are computed and the best combination of coefficients is found. Experimental results show that more coefficients are correctly quantized than the conventional direct requantization, and thus the PSNR as well as the subjective quality is much improved.