Catálogo de publicaciones - libros

The purpose of this paper is to visualize a still image and video on 3-dimensional medical image on the web using volume rendering board and VRML(Virtual Reality Modeling Language). Using VRML, the compressed 3-dimensional still image was represented effectively on the web through polygon reduction. The real time 3-dimensional volume rendering was compressed by Window Media Video 9 codec for implementation on the web. This paper focuses on detecting the most optimized 3-dimensional medical image on different bandwidths. Performance test has been carried out by evaluating the 3-dimensional medical image using various encoding bit rates on several network conditions. The network conditions were ADSL(Asymmetric Digital Subscriber Line), VDSL(Very high-data rate Digital Subscriber Line), LAN(Local Area Network), and Wireless LAN, which are common in South Korea.

We propose a novel retrieval method for fractal coded images in the compressed data domain. A fractal code is a contractive affine mapping that represents a similarity relation between two regions in an image. A fractal coded image consists of a set of these contractive mappings. Each mapping can be approximately represented by a vector spanning two regions. Therefore, a fractal coded image can be approximated as a set of vectors. By introducing a new similarity measure that reflects the difference of distribution and cardinality between two vector sets, a novel retrieval method for fractal coded images is realized. We also propose a new efficient retrieval method using upper bounds of the similarity measure. The effectiveness of the proposed method is also illustrated by various experiments.

An effective and robust image enhancement algorithm is presented for improving the visual quality of digital images captured under extremely low or non-uniform lighting conditions. The proposed algorithm is composed of two separated processes viz. adaptive luminance enhancement and adaptive contrast enhancement to provide a more flexible and better control over the image enhancement. Adaptive luminance enhancement is an intensity transformation based on a specifically designed nonlinear transfer function which largely increases the luminance of darker pixels and compresses the dynamic range as well. Adaptive contrast enhancement adjusts the intensity of each pixel based on its relative magnitude with respect to the neighboring pixels. Both processes can be self-tuned by the image statistical information. A proportional color restoration process is applied to convert the enhanced intensity image back to a color image. Real time processing and embedded application in mobile device have been successfully realized.

Although a lot of studies have been made on mobile learning, the study of content-based image recycling on mobile device is not known very well. This paper presents a new approach which recycles and augments existing off-line contents using a camera-equipped mobile device. Each individual learner has a PDA and an off-line textbook (Picture English Book: PEB). During the PEB-watching learning activity, users are dynamically provided with on-line information such as texts, videos and audios corresponding to the off-line contents via the PDA. A content-based image retrieval system (CBIR) is constructed to provide learner with required information using image recognition and multimedia technologies, such that the objective of m-learning can be achieved. We believe that it is worth developing a mobile learning system to provide the learners with a new educational environment which can recycles the existing PEBs.

We have generalised a class of similarity measures that are designed to address the problems associated with indexing high-dimensional feature space. The features are stored and indexed component wise. For each dimension we retrieve only those objects close the query point and then apply a local distance function to this subset. Thus we can dramatically reduce the amount of data looked at. We have evaluated these distance measures within a content-based image retrieval (CBIR) framework to determine the trade-off between the percentage of the data retrieved and the precision. Our results show that up to 90% of the data can be ignored whilst maintaining, and in some cases improving, retrieval performance.

Object movie refers to a set of images captured from different perspectives around a 3D object. Object movie is a good representation of a physical object because it can provide 3D interactive viewing effect, but does not require 3D reconstruction. In order to retrieve the desired object movie from the database, we first map an object movie into a manifold in the feature space. Two different sets of feature descriptors, one dense and one condensed, are designed to sample the manifold. Based on these descriptors, we define the dissimilarity measure between the query and the target in the object movie database. The query we considered can be either a complete object movie or simply a subset of views. In this paper, we further propose a relevance feedback approach to improving retrieved results. Some experimental results are shown to show the potential of our approach.

The development and use of content-based retrieval techniques for 3-D models is a relatively new departure in multimedia retrieval. We have extended our existing multimedia museum information system to support content-, metadata- and concept-based retrieval of 3-D models of museum artifacts and in this paper we describe a “classifier agent” to automatically assign associations between 3-D artifacts and concepts and metadata stored in a domain ontology. The context of the classifier agent is described, together with an overview of its architecture. Selecting appropriate parameters for the agent is an important activity and a comparison is made between manually selected parameters and the results of an automatic technique to determine “optimal” settings.

Accurate and robust correspondence calculations are the pre-requisite step in many 3D model query and retrieval process. However, the correspondence problem is particularly difficult for 3D biomedical model surfaces, especially for roundish and approximate symmetric organs such as liver, stomach, kidney etc. In this paper, we define a new feature representation called the Neighborhood Relative Angle context Distribution (NRACD) for each vertex and, based upon it, we apply the Chi-Square Goodness-of-Fit test to establish 3D point correspondence. We further define the similarities between correspondence ready models by Chi-Square test statistic values. The experimental results demonstrate that this approach is efficient and robust for surface point matching and is particularly applicable to the retrieval and analysis of 3D deformable objects.

The need for tools that effectively filter and efficiently search through a large amount of visual data is on the increase due to the rapid growth of multimedia information. Towards this goal, we propose a novel approach for image retrieval based on edge structural features using edge correlogram and color coherence vector. After color vector angle is applied to an image in the pre-processing stage, it is divided into two parts, as either smooth or edge pixels by the pixel classification. For the smooth pixels, the global color distribution of pixels is extracted by color coherence vector, incorporating spatial information into the proposed color descriptor. Meanwhile, for the edge pixels, the distribution of the gray pairs at an edge is extracted by edge correlogram. As the proposed method has both information for the local spatial correlation and information of global distribution of colors, it can be used to reduce the effect of the significant change in appearance and shape of objects. From the image representation based on edge structural features, the proposed algorithm provides a concise and flexible description even for the image with the complicated scenes. Experimental evidence shows that our algorithm outperforms the recent histogram refinement methods for image indexing and retrieval.

We investigate and compare the performance of several distributional distances in generic color image retrieval with an emphasis on symmetry and boundedness of the distances. Two histogram generation methods based on Gauss mixture vector quantization (GMVQ) are compared using Kullback-Leibler divergence (KLD). The joint histogram method shows a better retrieval performance than the Bayesian retrieval with the label histograms of interleaved data. A variety of distance measures are tested and compared for the joint histogram features produced by GMVQ, including an important set of Ali-Silvey distances, the Bhattacharyya distance, and a few other divergence measures based on Shannon entropy. Experimental results show that the Bhattacharyya distance and the L divergence are better than the histogram intersection (HI), but the KLD is poorer than the HI. In all cases, the symmetric version of a distance performs better than the asymmetric one and usually the bounded version of a distance gives better retrieval performance than the corresponding non-bounded.