Uwe Rauschenbach

Demand-driven image transmission with levels of detail and regions of interest

Abstract In this paper, we introduce a formal model which describes consistently levels of detail and regions of interest for the progressive transmission of raster images. This model serves as the foundation of a wavelet-based image communication framework which supports the progressive, redundancy-free transmission of multiple, possibly overlapping regions of interest in an image. To support demand-driven transmission, we introduce a method for interleaving control commands with the embedded bit stream. Since the dyadic wavelet decomposition allows the image resolutions in x and y direction to differ at most by a factor of two, we propose a new wavelet decomposition scheme which raises this number to a power of two for increased flexibility. We discuss some applications including the rectangular fish eye view, a focus and context display technique for raster images which efficiently uses this transmission method and saves screen real-estate.

The rectangular fish eye view as an efficient method for the transmission and display of large images

With the advances of mobile computing technology, the users of mobile hardware expect the same service as users of stationary computers do. Access to large images over the Internet poses two problems to users of wireless data communication devices and mobile computers: low transmission bandwidth and small screen space. In this paper, an integrated image transmission and display method is proposed, which efficiently uses both transmission bandwidth and screen real estate. In order to achieve that, we developed a level-of-detail and region-of-interest transmission scheme for raster images based on embedded zerotree wavelet coding. A special fish eye view technique is built on that transmission scheme, which can benefit directly from the properties of a modified wavelet decomposition.

General rectangular fisheye views for 2D graphics

Abstract This paper describes the R ectangular F ish E ye V iew for the combined presentation of 2D raster images, 2D vector graphics and text. A detailed presentation of the area of highest user interest (the focus) is integrated into a lower-detail context display providing an overview. Screen space is saved by downscaling (distorting) the context. The geometric layout of the view is assembled using rectangular regions. Three different context modes are proposed to provide for computational scalability. Performance and quality implications of two different implementation options are compared.

A presentation model for mobile information visualization

Abstract One of the visions of mobile computing is to put “all information at the users fingertips”—to allow a user to operate on any data, any time, anywhere. The idea is to create an information environment providing homogeneous access to all data and services available in the distributed, mobile computing infrastructure. A fundamental requirement for access to such an open, distributed information system is an intelligent selection of methods for information visualization based on user requirements and available display functionality. In this paper, a flexible concept is proposed that allows to enrich the nodes of an information structure with information about which alternative display methods can be used for what parts of the node. These ‘facets’ are then used by a recursive view generation process for selecting suitable display methods while creating a visualization of an information structure. Influence parameters such as user characteristics, display resources, and data properties can be used to guide the selection process in order to create a presentation that optimally meets the users goals.

Compression of palettized images with progressive coding of the color information

This paper introduces a new compression method for palettized images, which supports progressive refinement of the color information in contrast to the resolution refinement used in standard methods like interlaced GIF. Such, fine image details can be recognized after decoding only a small part of the compressed image data. Achieved compression ratios are comparable to those of interlaced GIF or PNG. The method combines color map sorting with bitplane by bitplane prediction and Golomb coding of the pixel field.

Video analysis for universal multimedia messaging

Multimedia messaging is expected to become a major application for next generation mobile computing and communication devices. Heterogeneous capabilities of these devices, however require adaptation of multimedia messages before rendering on any specific device. To achieve this in case of video messages, video analysis has to be performed to extract the structure of the video and control the message adaptation. This paper describes the universal multimedia messaging scenario and presents a short overview of existing video structuring approaches. A computationally efficient method for analyzing a video message in this scenario is introduced and a prototype for universal multimedia messaging conforming to the emerging MPEG-7 standard is outlined.

Wireless live streaming video of surgical operations: an evaluation of communication quality

We evaluated a mobile video system for surgical teleconsultation. A video streaming server in the operating room transmitted video and audio to a hand-held computer (personal digital assistant [PDA]) over a wireless local area network. Two groups of 20 surgeons (each with 12 qualified surgeons and eight surgeons between the 2nd and the 4th year of training) participated in the tests. For voice transmission, correct understanding of numbers was achieved in 100% of the cases (n = 1000) and 98% of medical terms (n = 400). The quality of the video displayed on the PDA was assessed by the recognition of different operating room scenarios. Only 62% (SD 17) of the structures were identified clearly on the hand-held device (n = 400). The accuracy improved to 78% (SD 15) (n = 400) if the same scenario was observed on a larger (50 cm) video screen (p < 0.001). Accuracy was significantly better if audio conversation was possible. The quality evaluation by the consultants showed that the PDA display size and quality were sufficient for clinical use.

Interactive TV: A new application for mobile computing

Abstract This contribution presents some recent developments in the field of interactive digital television and discusses the challenges of bringing these TV services to mobile devices. The paper addresses two different scenarios. The first part discusses the use case of complementing the TV set by mobile devices in a home network, considering also the aspect of using different networks (both broadcast and IP-based data networks) to deliver the content to the home. In the second part, the new field of mobile broadcast technologies is introduced and an overview over the different technologies and standardisation activities is provided.

Bayesian method for motion segmentation and tracking in compressed videos

This contribution presents a statistical method for segmentation and tracking of moving regions from the compressed videos. This technique is particularly efficient to analyse and track motion segments from the compression-oriented motion fields by using the Bayesian estimation framework. For each motion field, the algorithm initialises a partition that is subject to comparisons and associations with its tracking counterpart. Due to potential hypothesis incompatibility, the algorithm applies a conflict resolution technique to ensure that the partition inherits relevant characteristics from both hypotheses as far as possible. Each tracked region is further classified as a background or a foreground object based on an approximation of the logical mass, momentum, and impulse. The experiment has demonstrated promising results based on standard test sequences.

Temporal video segmentation using global motion estimation and discrete curve evolution

The identification of syntactic or semantic temporal segments is an important process of video-content analysis. The paper proposes a temporal video segmentation method based on global motion in order to analyze meaningful temporal substructures of camera shots. To ensure that the detected segment optimally contributes to the shot global characteristic, the proposed method exploits a state-of-the-art discrete curve evolution. This technique leads to a subdivision of the global motion trajectory, where each segment of the subdivision has a constant relevant global motion. Experimental results based on standard test sequences acknowledge the method functionality, especially for shots characterized by pronounced camera motion.