适用于视觉媒体检索的视频镜头分割算法*

针对基于内容的视觉媒体检索存在的问题,依据多维空间仿生信息学（multi-dimensional space biomimetic informatics,MDSBI）理论的同源连续性规律（principle of homology continuity,PHC）,在高维空间中通过几何方法分析视频及镜头和图像,通过子空间方法解决不同视觉媒体的异构性问题,得出视觉媒体检索的实质性工作是在高维空间研究点和子空间的距离问题,并实现适用于视觉媒体检索的镜头分割算法。     

结合视觉语义的乳腺X线图像检索

针对乳腺X线图像检索过程中底层特征不能有效表达高层语义,提出一种结合视觉语义的图像检索算法.引入图半监督学习框架提取查询图像的类别归属度,作为视觉语义;考虑类别分布的不平衡性,在语义提取过程中加入类标签正则化;设计结合视觉语义和底层特征的相似度度量准则.在乳腺X线分块图像数据库上的实验结果表明,该算法能够有效提取图像的视觉语义,性能优于基于单一底层特征的检索算法.     

综合颜色纹理形状特征的图像检索

图像特征的提取和使用在基于内容的图像检索中至关重要.研究了在基于内容的图像检索系统中整合颜色,纹理,形状的提取方法.将图像按照一定的规则进行分块,对各个分块分别进行各种特征向量的提取.颜色特征的提取是基于YUV颜色空间的颜色直方图,纹理特征的提取采用Gabor滤波器,形状特征的提取是基于Zernike矩的计算.实验结果表明,综合图像的颜色、形状和纹理特征提高了图像检索的准确性.     

Access to Pictorial Material: A Review of Current Research and Future Prospects

Rapid expansion in the digitization of image and image collections has vastly increased the numbers of images available to scholars and researchers through electronic means. This research review will familiarize the reader with current research applicable to the development of image retrieval systems and provides additional material for exploring the topic further, both in print and online. The discussion will cover several broad areas, among them classification and indexing systems used for describing image collections and research initiatives into image access focusing on image attributes, users, queries, tasks, and cognitive aspects of searching. Prospects for the future of image access, including an outline of future research initiatives, are discussed. Further research in each of these areas will provide basic data which will inform and enrich image access system design and will hopefully provide a richer, more flexible, and satisfactory environment for searching for and discovering images. Harnessing the true power of the digital image environment will only be possible when image retrieval systems are coherently designed from principles derived from the fullest range of applicable disciplines, rather than from isolated or fragmented perspectives.     

一种新型的音乐信息检索索引方法及其应用

本文提出一种新型的音乐信息检索索引方法,并将其应用于哼唱检索系统上,获得可喜的成果。该方法根据音乐旋律的特点,使用音高和音长描述一段乐曲。考虑到各种音乐输入方式存在的音高不一致性和节奏伸缩性,系统根据乐曲旋律特征的总体变化,将平均音高差和平均音长变化百分比作为关键字建立M树索引结构。由于各类输入旋律常常存在误差,本文定义一定的搜索半径,查找半径范围内与输入的音乐片断最相似的音乐文件。实验结果表明,该结构可以使系统提高60%以上的检索速度,并且搜索结果的正确率最高仍然可达80%。     

基于内容的MRI脑肿瘤图像特征提取及检索方法

医学图像检索中特征提取方法对检索的效果、性能具有重要影响,针对这个问题,设计了一个基于内容的医学图像检索系统.为了给医学图像检索系统的临床应用提供参考价值,该系统以哈佛大学医学院开发的脑肿瘤MRI医学图像数据库为背景,比较了颜色相关图、颜色矩、灰度共生矩阵、金字塔小波变换和树型小波变换这5个特征提取技术对MRI脑肿瘤医学图像的检索性能.实验结果表明树型小波变换和金字塔小波变换的检索效果较好.     

一种因果的突变镜头检测方法

通常镜头检测算法在进行镜头检测时都要利用前后若干帧的信息,该文提出了一种因果的突变镜头检测方法,只利用当前帧之前有限长度内帧的信息。利用两帧之间3×3分块后对应区间HSV直方图间的距离,应用加权系数来计算帧间差值。运用位于当前帧前的一个可变长的滑动窗口来进行突变镜头检测,并对静止镜头进行了处理。该文最后给出了算法的完整流程并进行了试验,试验结果表明该算法的检测准确率达到了95%,并对推拉镜头具有一定的鲁棒性。     

A fractal-based clustering approach in large visual database systems

Large visual database systems require effective and efficient ways of indexing and accessing visual data on the basis of content. In this process, significant features must first be extracted from image data in their pixel format. These features must then be classified and indexed to assist efficient access to image content. With the large volume of visual data stored in a visual database, image classification is a critical step to achieve efficient indexing and retrieval. In this paper, we investigate an effective approach to the clustering of image data based on the technique of fractal image coding, a method first introduced in conjunction with fractal image compression technique. A joint fractal coding technique, applicable to pairs of images, is used to determine the degree of their similarity. Images in a visual database can be categorized in clusters on the basis of their similarity to a set of iconic images. Classification metrics are proposed for the measurement of the extent of similarity among images. By experimenting on a large set of texture and natural images, we demonstrate the applicability of these metrics and the proposed clustering technique to various visual database applications.     

图像主题区域提取及其在图像检索中的应用

图像主题区域是表达图像主要语义内容的部分.根据图像不同区域的视觉显著性差异特点,利用改进的视觉注意模型,提取图像的主题区域,并划分为不同语义类别.提出基于主题区域的图像检索方法,该方法能按照图像主题区域所属的语义类别,在图像之间有针对性地选择同类别主题区域进行区域相似性比较.实验结果表明,此方法可以更有效地表达图像的主要语义内容,提高图像检索的效率和准确性.