EGMM04: EG Multimedia Workshop 2004

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/351

Browse

Now showing 1 - 20 of 20

Analysis of Inter-Frame Coding Without Intra Modes in H.264
(The Eurographics Association, 2004) Cheng, Yun; Wang, Zhiying; Dai, Kui; Guo, Jianjun; N. Correia and J. Jorge and T. Chambel and Z. Pan
H.264/AVC is a new international standard for video coding which has great advantage of coding efficiency com-pared with other standards. It can save about 50% bit-rate compared with that of the successful prior coding stan-dards under the same reconstructed picture quality. But the high coding efficiency is acquired by heavily computa-tion. In this paper, the coding mode and algorithm for mode decision are introduced firstly, then transform and quantization are analyzed and experiments on inter-frame coding with or without intra modes are performed. The experiment results illustrate that the encoding method without intra modes in inter-frame coding will decrease the encoding time from 76.03% to 50.09% compared with that of the standard encoding method, while the PSNR-Y will change from -0.45dB to +0.20dB (most cases are ±0.10dB) at the same bit-rates.
Progressive Image Transmission Using Singular Value Decomposition
(The Eurographics Association, 2004) Chang, Chin-Chen; Liu, Yi-Long; N. Correia and J. Jorge and T. Chambel and Z. Pan
There are many progressive image transmission schemes using vector quantization-related schemes, but there are some problems associated with them. First, these schemes need to train a codebook, and this codebook will directly decide the quality of the recovered image. VQ-related PIT methods also need some time to search the codebook to find the indices of corresponding vectors. Thus, this paper looks for a new PIT method without VQ. By applying SVD to PIT, image will be decomposed into three matrices. After decomposition, certain processes are applied to these matrices, and PIT will be achieved by transmitting these three matrices. The use of the proposed method resulted in higher image quality than would result from the use of traditional methods, like the bit-plane method and the improved bit-plane method, in the experiments conducted in this paper. This paper proposes a new way of PIT, and this new method is achieved PIT without VQ.
A Chinese Remainder Theorem Oriented Information Hiding Scheme
(The Eurographics Association, 2004) Chang, Chin-Chen; Lu, Tzu-Chuen; N. Correia and J. Jorge and T. Chambel and Z. Pan
Steganography is an information hiding technique that conveys secret information in a host signal using a secret method. Only the receivers and senders know the secret information. Many researchers have proposed their own steganographic techniques to hide information in various host signals, such as audios, videos, images, and so on. Nevertheless, most of the methods degrade the visual quality of the image when more information is hidden in the image. Therefore, this paper proposes a new steganographic scheme, which is based on the Chinese Remainder Theorem. The abbreviation of the scheme is CRTIH, and it not only conceals a larger amount of information in a hidden image but also upgrades the visual quality of the image.
A Component-based Authoring Environment for Creating Multimedia-RichMixed Reality
(The Eurographics Association, 2004) Abawi, Daniel; Dörner, Ralf; Grimm, Paul; N. Correia and J. Jorge and T. Chambel and Z. Pan
Applications that seek to combine multimedia with Mixed Reality (MR) technologies in order to create multimediarich MR environments pose a challenge to authors who need to provide content for such applications. Founded on a component-based authoring paradigm, production processes as well as tools that serve as a supportive authoring environment for these authors are presented in this paper. For this, not only requirements that stem from multimedia authoring or MR authoring alone have been identified but also authoring tasks that are only present in the creation of multimedia-rich MR content. Concepts for supporting these tasks within a componentbased authoring framework (e.g. the specification of phantom objects) are presented. The resulting authoring tools are discussed - one of their main advantages is that they provide a direct preview of the content for the author. This allows multimedia authors who are not familiar with MR methodologies to quickly gain experience with multimedia-rich MR content creation.
Content Based Image Public Watermarking
(The Eurographics Association, 2004) Hengfu, Yang; Zihua, Yang; Mingfang, Jiang; N. Correia and J. Jorge and T. Chambel and Z. Pan
In this paper, a content based image public watermarking technique which operates in DCT domain is proposed. First, the 8×8 DCT sub-blocks of the host image are rearranged into a Hilbert sequence in Hilbert scanning order, then two neighboring sub-blocks in the Hilbert sequence is pseudo-randomly selected by using chaotic sequences. Then a watermark with visually recognizable pattern is embedded into the original image by changing the polarity of the corresponding middle-frequency coefficients in the two chosen neighboring sub-blocks, and the watermark is adapted to the image by exploiting the masking characteristics of the human visual system (HVS), thus ensuring the watermark invisibility, and the watermark don't need the original image. The experimental results show that the proposed algorithm in this paper is robust to common signal processing techniques and some geometric distortions, such as cropping, scaling and rotation. Especially, it achieves high robustness under signal enhancement operations, such as sharpening, contrast enhancement, edge enhancement and histogram equalization.
Optimum Detection of MultiplicativeWatermarks for Digital Images in the DWT Domain
(The Eurographics Association, 2004) Sun, Zhongwei; Feng, Dengguo; Xue, Rui; N. Correia and J. Jorge and T. Chambel and Z. Pan
Watermark detection plays a crucial role in digital watermarking. It has traditionally been tackled using correla-tion-based techniques. However, the correlation-based detection is not the optimum choice when the host media doesn t follow a gaussian distribution or the watermark is not embedded in the host media in an additive way. A discrete wavelet transform (DWT) domain multiplicative watermark detection algorithm for digital images is propo-sed in this paper, which exploits the imperceptibility constraint of watermarking. By formulating the watermark detection as weak signal detection in non-gaussian noise, the proposed algorithm is derived according to statistical inference theory. With the wavelet coefficients modeled by generalized gaussian distribution (GGD), the optimum decision threshold for the detector is obtained by applying Neyman-pearson criteria. The superiority of the novel detector in performance is confirmed through Monte Carlo simulations. Keywords: Digital watermarking, Multiplicative embedding, Discrete wavelet transform, Generalized gaussian dis-tribution, Weak signal detection.
A Web Services-Based Architecture for Capability-Aware Ubiquitous Media
(The Eurographics Association, 2004) Haipeng, Wang; Xingshe, Zhou; Zongtao, Duan; Tao, Zhang; N. Correia and J. Jorge and T. Chambel and Z. Pan
Ubiquitous media aims to provide media services anytime and anywhere. To realize it, one challenge is how to provide customized and dynamic services to a variety of computing devices with different capabilities. This paper presents an architecture, which enables customized delivery of multimedia services. A concept of capability is used to abstract the adaptation-related attributes of computing devices. Capability concerns both the static and dynamic attributes of the computing devices, such as display resolution and remaining battery power. These two classes of attributes can be combined to provide complementary information for customized and dynamic media delivery. As a proof of concept, we have developed a prototype implementation, which is characterized by Web services-based architecture, and capability-aware feature. Our initial experiments show the effectiveness of this architecture.
A Segmentation Algorithmfor Jacquard Images Based on Mumford-ShahModel
(The Eurographics Association, 2004) Feng, Z. L.; Yin, J. W.; Chen, G.; Liu, Yang; Dong, J. X.; N. Correia and J. Jorge and T. Chambel and Z. Pan
Automatic pattern segmentation of jacquard images is a challenging task due to the complexity of the images. Active contour models have become popular for finding the contours of a pattern with a complex shape. However, these models have many limitations on the pattern segmentation of jacquard images in the presence of noise. In this paper, a robust algorithm based on the Mumford-Shah model is proposed for the segmentation of noisy jacquard images. We discretize the Mumford-Shah model on piecewise lin-ear finite element spaces to yield greater stability and higher accuracy. A novel iterative relaxation algo-rithm for the numerical solving of the discrete version of the Mumford-Shah model is presented. During each iteration, we first refine and reorganize an adaptive triangular mesh to characterize the essential contour structure of a pattern. Then, we apply the quasi-Newton algorithm to find the absolute minimum of the discrete version of the model at the current iteration. Experimental results on synthetic and jac-quard images have shown the effectiveness and robustness of the algorithm.
A Novel Quadtree-Structured Scheme for Transmitting Chinese Calligraphy Progressively
(The Eurographics Association, 2004) Chang, Chin-Chen; Li, Chien-Fa; Lin, Iuon-Chang; N. Correia and J. Jorge and T. Chambel and Z. Pan
Progressive image transmission (PIT) techniques are useful in Web applications. A user can get a rough preview of an image and decide if the entire detail of a picture needs to be transmitted, especially when the network bandwidth is restricted. In this paper, we propose a new progressive image transmission scheme for Chinese calligraphy. Because the colors used in Chinese calligraphy are very simple, we can apply the quadtree-structure to transmit the calligraphy progressively. The preliminary results show that transmitted bits can be saved in the first few rounds. The PSNR values in the first few rounds are much better than other PIT techniques.
Mixing Images and Sketches for Retrieving Vector Drawings
(The Eurographics Association, 2004) Ferreira, A.; Fonseca, M. J.; Jorge, Joaquim A.; Ramalho, M.; N. Correia and J. Jorge and T. Chambel and Z. Pan
Current approaches to content-based retrieval of multimedia data usually rely either on query by example or on sketches of the desired image, but not on both. In this paper, we propose a new query specification scheme, where digital images are combined with sketches, after vectorization, taking advantage of both methods. We selected a set of algorithms to perform image vectorization, taking into account the trade-off between vector image quality and processing time. This method of specifying queries is part of a system to retrieve vector drawings, which we briefly describe in this paper.
Remote Raster Image Browsing Based on Fast Content Reduction forMobile Environments
(The Eurographics Association, 2004) Rosenbaum, René; Schumann, Heidrun; N. Correia and J. Jorge and T. Chambel and Z. Pan
Enhanced browsing techniques for digital imagery and small displays facilitate the exploration process of large images often by using new ways to represent the image. Reduction of image content is such an approach mostly linked with need for strong processing power. To overcome this, we propose the use of the Discrete Wavelet Transform (DWT), which inherently separates detail and approximation of the image. By enhancing the detail and removing the approximation directly in wavelet domain, a very fast content reduction can be achieved. Due to its flexibility, JPEG2000 is used as basis for an efficient system for remote image browsing. To satisfy demands which are imposed by the use of current mobile hardware, every stage of the image communication pipeline is adapted and tightly coupled to the used browsing technique to reduce the need for processing power and bandwidth.
Door Access Control Using Human Face and Height
(The Eurographics Association, 2004) Zhang, H. X.; Ma, R. H.; Huang, W. M.; Huang, Z.; N. Correia and J. Jorge and T. Chambel and Z. Pan
Access control has much attracted research interest recently. In this paper, we propose a method using human face and height as human trait to recognize a person. We observe that eye location extracted from a human face is sta-ble to be used to compute his/her height to the ground. Using it together with face recognition can increase the ac-curacy of the access control. We have implemented the method on a PC installed with a stereo camera. The design criteria, techniques, implementation details, and performance testing are presented. Keywords: 3D reconstruction, gradient descent method, biometric fusion, application.
Simultaneous Tracking of Multiple Objects for Augmented Reality Applications
(The Eurographics Association, 2004) Yuan, C.; N. Correia and J. Jorge and T. Chambel and Z. Pan
This paper presents an appearance-based image processing and tracking algorithm which is applied in a distributed Augmented Reality (AR) system. The tracker is computer vision based and is capable of simultaneous tracking of multiple objects. These objects are called place holder objects (PHOs), as they are used as interface elements and act as tangible interfaces for handling and interacting with virtual artifacts. The tracking system uses a fix mounted camera viewing at the workspace - a normal round table. All the PHOs are placed on the table and can be moved arbitrarily around, allowing both in-plane and out-of-plane rotations. In order to track and differentiate the PHOs in real-time, we apply an appearance-based object modeling. The utilization of appearance-based method for object recognition and tracking gives the system a distinct advantage in that it is computationally less expensive and it can be easily adapted to work with arbitrary PHOs by simply using an off-line training process.
Estimating Traffic Density Using Sounds of Moving Vehicles
(The Eurographics Association, 2004) Kato, Jien; Hiramatsu, Yoshitaka; Watanabe, Toyohide; N. Correia and J. Jorge and T. Chambel and Z. Pan
This paper proposed a method for automatically estimating traffic density by using sounds of moving vehicles. The approach is based on the idea of recognition of the temporal variations that appear on the power signals when vehicles pass through an observation point. The local temporal variations in small periods of time are extracted by wavelet transformation and are used as an observation sequence for a hidden Markov model, which models the global temporal variations of the power signal. The passages of vehicles are detected based on the state transitions of the HMM. The occlusion problem due to the overlapping of the sounds of moving vehicles are dealt with by corresponding two set of information from a stereo microphone. Experimental results show that with some restrictions, the passages of vehicles are able to be detected from road traffic sounds in good accuracy, by the proposed method.
Image-based Rendering of the AnisotropicBRDF ofWoven Fabrics
(The Eurographics Association, 2004) Takeda, Yuki; Viet, Huynh Quang Huy; Tanaka, Hiromi T.; N. Correia and J. Jorge and T. Chambel and Z. Pan
The reflectance of fabric surface is commonly represented by a 4D bidirectional reflectance distribution function (BRDF). To generate the BRDF from measured data by a gonioreflectometer with 2 degrees of freedom of the light source and 2 degrees of freedom of the observing direction, it requires an enormous amount of measurements. In this paper, we propose an efficient image-based method for rendering the anisotropic BRDF of woven fabrics based on the micro facet surface geometry determined by the cross-sectional shape of fibers, twist of yarns, and type of weave. At first, we examine the relationship between the reflectance properties and the micro facet surface geometry of a type of woven fabric such as silk-like synthesized fabric. Next, we develop an image-based method for generating the BRDF of woven fabrics from measurement of the reflectances caused by the incident light only in the direction perpendicular to the fabric s surface. The simulation results on arbitrarily colored dresses show the performance of the proposed approach.
Animating Peer-Level Annotations Within Web-Based Multimedia
(The Eurographics Association, 2004) Bulterman, Dick; N. Correia and J. Jorge and T. Chambel and Z. Pan
The TabletPC is an example of a new generation of user interface device where pen-based manipulation of information is integrated directly into a user s workflow. Using the TabletPC's existing pen and electronic ink systems, a wide range of static documents can be created or annotated. While the facilities of the TabletPC are useful for creating virtual images containing ink that can be overlaid on text or picture context, there is little support for creating annotations of time-based content such as video. This article describes an annotation authoring model and interface for creating peer-level annotations to video media. Peer-level annotations allow existing content to be enriched with additional content annotations that can be co-presented with the original media. A system for creating a SMIL language document containing SVG-based annotations that exist along-side the visual content is described, along with a discussion of the needs and limitations of supporting video markup in a web context. An example using peer-level annotations in a medical context is provided.
Tetrahedral Adaptive Grid for Parallel Hierarchical Tetrahedrization
(The Eurographics Association, 2004) Takama, Yasufumi; Kimura, Akinori; Tanaka, Hiromi T.; N. Correia and J. Jorge and T. Chambel and Z. Pan
Recent advances in volume scanning techniques have made the task of acquiring volume data of 3-D objects easier and more accurate. Since the quantity of such acquired data is generally very large, a volume model capable of compressing data while maintaining a specified accuracy is required. The objective of this work is to construct a multi-resolution tetrahedral representation of input volume data. This representation adapts to local field properties while preserving their discontinuities. In this paper, we present an accuracy-based adaptive sampling technique to construct a multi-resolution model, we call a tetrahedral adaptive grid, for hierarchical tetrahedrization ofC1 continuous volume data.We have developed a parallel algorithm of tetrahedral adaptive grid generation that recursively bisects tetrahedral gird elements by increasing the number of grid nodes, according to local field properties and such as orientation and curvature of isosurfaces, until the entire volume has been approximated within a specified level of view-invariant accuracy. We have also developed a parallel algorithm that detects and preserves both C0 and C1 discontinuities of field values, without the formation of cracks which normally occur during independent subdivision. Experimental results obtained using a PC cluster system demonstrate the validity and effectiveness of the proposed approach.
A New Approach to Multimedia Information Filtering Based on its Structure
(The Eurographics Association, 2004) Huang, Xiaodi; Yong, Jianming; N. Correia and J. Jorge and T. Chambel and Z. Pan
In information filtering systems, the multimedia documents are sequentially presented to users based on the user relevance values. This paper argues that the presented multimedia documents should be both important and relevant to the users. The importance of a document is determined by its relations to others in the collection. All users are supposed to look for important and relevant documents. Based on this view, a structure-based filtering framework is described, which incorporates the characteristics of the importance and relevance of multimedia documents. An approach to calculating importance values of multimedia documents and then combining them into relevance values of multimedia documents is proposed to improve the representation of user profiles. An example is provided.
A Training-Based Method for Reducing Ringing Artifact in BDCT-Encoded Images
(The Eurographics Association, 2004) Wang, Guangyu; Wong, Tien-Tsin; Heng, Pheng-Ann; N. Correia and J. Jorge and T. Chambel and Z. Pan
The quantization procedure of block-based discrete cosine transform (BDCT) compression (such as JPEG) introduces annoying visual artifact. In this paper, we propose a novel training-based method to reduce the ringing artifact in BDCT-encoded high-contrast images (images with large smooth color areas and strong edges/outlines). Our main focus is on the removal of ringing artifact that is seldom addressed by existing methods. In the proposed method, the contaminated image is modeled as a Markov random field (MRF). We learn the behavior of contamination by extracting massive number of artifact patterns from a training set. To organize the extracted artifact patterns, we use the tree-structured vector quantization (TSVQ). Instead of post-filtering the input contaminated image, we synthesize an artifact-reduced image. We show that substantial improvement (both statistical and visual) is achieved using the proposed method. Moreover, since our method is non-iterative, it can remove artifact within a very short period of time.
A New Fingerprint Image Segmentation Algorithm Based on ROIO
(The Eurographics Association, 2004) Shi, Zhongchao; Xu, Ke; Wang, Yangsheng; N. Correia and J. Jorge and T. Chambel and Z. Pan
We present a new fast and effective method of fingerprint image segmentation, which is different from traditio-nal methods that usually use some certain features to segment diversified images. Based on the region of inte-rest, an approach is introduce, which is specialized on preprocessing a certain kind of bad fingerprint images with a great many blurs, false traces and sweat spots. It just mimics the way of finding region of interest by human beings and locating the area of low intensities quickly. Experimental result shows a significant improve-ment in fingerprint segmentation performance.

Browse

Browsing EGMM04: EG Multimedia Workshop 2004 by Issue Date

Results Per Page

Sort Options