Querying on large and complex databases by content: Challenges on variety and veracity regarding real applications

The amount and variety of digital data currently being generated, stored and analyzed, including images, videos, and time series, have brought challenges to data administrators, analysts and developers, who struggle to comply with the expectations of both data owners and end users. The majority of the applications demand searching complex data by taking advantage of queries that analyze different aspects of the data, and need the answers in a timely manner. Content-based similarity retrieval techniques are well-suited to handle large databases, because they enable performing queries and analyses using features automatically extracted from the data, without users’ intervention. In this paper, we review and discuss the challenges posed to the database and related communities in order to provide techniques and tools that can meet the variety and veracity characteristics of big and complex data, while also considering the aspects of semantical preservation and completeness of the data. Examples and results obtained over a two-decade-long experience with real applications are presented and discussed.     

Texture image retrieval using adaptive tetrolet transforms

This paper proposes a novel technique for texture image retrieval based on tetrolet transforms. Tetrolets provide fine texture information due to its different way of analysis. Tetrominoes are applied at each decomposition level of an image and best combination of tetrominoes is selected, which better shows the geometry of an image at each level. All three high pass components of the decomposed image at each level are used as input values for feature extraction. A feature vector is created by taking standard deviation in combination with energy at each subband. Retrieval performance in terms of accuracy is tested on group of texture images taken from benchmark databases: Brodatz and VisTex. Experimental results indicate that the proposed method achieves 78.80% retrieval accuracy on group of texture images D1 (taken from Brodatz), 84.41% on group D2 (taken from VisTex) and 77.41% on rotated texture image group D3 (rotated images from Brodatz).     

A new SVM-based relevance feedback image retrieval using probabilistic feature and weighted kernel function

Relevance feedback (RF) is an effective approach to bridge the gap between low-level visual features and high-level semantic meanings in content-based image retrieval (CBIR). The support vector machine (SVM) based RF mechanisms have been used in different fields of image retrieval, but they often treat all positive and negative feedback samples equally, which will inevitably degrade the effectiveness of SVM-based RF approaches for CBIR. In fact, positive and negative feedback samples, different positive feedback samples, and different negative feedback samples all always have distinct properties. Moreover, each feedback interaction process is usually tedious and time-consuming because of complex visual features, so if too many times of iteration of feedback are asked, users may be impatient to interact with the CBIR system. To overcome the above limitations, we propose a new SVM-based RF approach using probabilistic feature and weighted kernel function in this paper. Firstly, the probabilistic features of each image are extracted by using principal components analysis (PCA) and the adapted Gaussian mixture models (AGMM) based dimension reduction, and the similarity is computed by employing Kullback–Leibler divergence. Secondly, the positive feedback samples and negative feedback samples are marked, and all feedback samples’ weight values are computed by utilizing the samples-based Relief feature weighting. Finally, the SVM kernel function is modified dynamically according to the feedback samples’ weight values. Extensive simulations on large databases show that the proposed algorithm is significantly more effective than the state-of-the-art approaches.     

改进相关反馈技术在CBVR人体动作识别中的应用研究

针对CBVR人体动作识别问题,提出利用相关反馈技术与其它技术相结合。首先,分析了从视频中提取人类行为并将其表示成特征集的方法;然后研究了各种基于内容的视频检索方法;最后,将这些技术进行组合得到基于相关反馈技术的最优CBVR人体动作识别方法。在三个行为数据库中(包括UCF运动,UCF视频网站和HOHA2)评估了由上述几种技术组合而成的性能,实验结果为开发高效的基于内容的视频检索系统提供了有益的借鉴。     

Content-based assembly search: A step towards assembly reuse

The increased use of CAD systems by product development organizations has resulted in the creation of large databases of assemblies. This explosion of assembly data is likely to continue in the future. In many situations, a text-based search alone may not be sufficient to search for assemblies and it may be desirable to search for assemblies based on the content of the assembly models. The ability to perform content-based searches on these databases is expected to help the designers in the following two ways. First, it can facilitate the reuse of existing assembly designs, thereby reducing the design time. Second, a lot of useful designs for manufacturing, and assembly knowledge are implicitly embedded in existing assemblies. Therefore a capability to locate existing assemblies and examine them can be used as a learning tool by designers to learn from the existing assembly designs. This paper describes a system for performing content-based searches on assembly databases. We identify templates for comprehensive search definitions and describe algorithms to perform content-based searches for mechanical assemblies. We also illustrate the capabilities of our system through several examples.     

Content-based image retrieval using H.264 intra coding features

Efficient multimedia retrieval has become a vital issue because more audio and video data are now available. This paper focuses on content-based image retrieval (CBIR) in the compression domain (CPD). The retrieval features are extracted based on I-frame coding information in H.264. This paper proposes using a local mode histogram as the texture feature to match images and applying the residual coefficients to filter non-confident modes. The geometrical correspondence between two images is also considered. The experimental results show that the proposed method can substantially reduce computational and memory resource consumption, and provides similar performance compared with methods that extract features from decompressed images.     

用CORBA技术构建分布式图像检索系统

首先针对现有Web图像搜索引擎设计的局限性 ,提出了分布式系统集成的重要性。然后 ,对CORBA技术进行了详细的描述 ,说明了CORBA技术是构建分布式图像检索系统的理想平台 ,并且介绍了图像检索技术。最后 ,提出了一种基于CORBA技术的分布式图像检索系统设计模型 ,并详细描述了其体系结构。     

Zernike矩不变性分析及其改进

在分析了Zernike矩不变性后,提出了一种基于图像整个区域形状描述算子的改进Zernike矩。改进的Zernike矩不仅具有旋转不变性,而且还具有改进前不具有的比例不变性。试验结果证明,和改进前的Zernike矩相比,改进后的Zernike矩具有更好的不变特征,可用作基于区域的形状描述算子。