Watermarking is an emerging technology that is claimed to have an important application in copyright protection. In this paper we consider the case of video watermarking for DVD-Video copy protection. An overview is given of the system requirements and its implications for the DVD-Video watermarking technology. In particular we focus on the problem of implementing copy generation control. Finally, we report on the status of the standardization discussions taking place in the Copy Protection Technical Working Group. Part of this paper has been presented at the 1999 edition of the ICMCS.

Previous works on data hiding generally targeted on a specific tradeoff between capacity and robustness. This results in overestimation of the processing noise under some situations and/or underestimation under some other situations, hence limits the overall performance. In this paper, we propose a multi-level data hiding scheme which is able to convey secondary data in high rate when noise is not severe and can also convey some data reliably under heavy distortion. The proposed scheme is motivated by a two- category classification of embedding schemes and by a study on detection performance of spread spectrum watermarking. The multi- level data hiding has been successfully applied to both digital image and video, and can be used for applications such as copy control.

The use of digital images and video is growing on the Internet and on consumer devices. Digital images and video are easy to manipulate, but this ease of manipulation makes tampering with digital content possible. Examples of the misuse of digital content include violating copyrights of the content and tampering with important material such as contents of video surveillance. In this paper we present an algorithm that extracts a binary signature from an image. This approach can be used to search for possible copyright violations by finding images with signatures close to that of a given image. The experimental results show that the algorithm can be very effective in helping users to retrieve sets of almost identical images from large collections of images. The signature can also be used for tamper detection. We will show that the signatures we extract are immune to quantization errors that result from compression and decompression.

In this paper, we concentrate on the problem of robust watermarking in the presence of perceptual coding. We first present a watermarking approach called Robust Reference Watermarking (RRW) which can be incorporated into a broad class of watermarking algorithms to improve their performance. Through analysis of this scheme we demonstrate how embedding the watermark using a different perceptual model and domain than for compression can result in improved watermark extraction and detection reliability. Simulation results are also presented to verify our theoretical observations.

Copyright notification and enforcement, authentication, covert communication, and hybrid transmission are examples of emerging multimedia applications for digital watermarking methods, methods for embedding one signal (e.g., the digital watermark) within another 'host' signal to form a third, 'composite' signal. The embedding is designed to achieve efficient trade-offs among the three conflicting goals of maximizing information-embedding rate, minimizing distortion between the host signal and composite signal, and maximizing the robustness of the embedding. Quantization index modulation (QIM) methods are a class of watermarking methods that achieve provably good rate-distortion-robustness performance. Indeed, QIM methods exist that achieve performance within a few dB of capacity in the case of a (possibly colored) Gaussian host signal and an additive (possibly colored) Gaussian noise channel. Also, QIM methods can achieve capacity with a type of postprocessing called distortion compensation. This capacity is independent of host signal statistics, and thus, contrary to popular belief, the information-embedding capacity when the host signal is not available at the decoder is the same as the case when the host signal is available at the decoder. A low-complexity realization of QIM called dither modulation has previously been proven to be better than both linear methods of spread spectrum and nonlinear methods of low-bit(s) modulation against square-error distortion-constrained intentional attacks. We introduce a new form of dither modulation called spread-transform dither modulation that retains these favorable performance characteristics while achieving better performance against other attacks such as JPEG compression.

Blind or oblivious data hiding, can be considered as a signaling method where the origin of the signal constellation is not known. The origin however, can be estimated, by means of self-noise suppression techniques. In this paper, we propose such a technique, and present both theoretical and numerical evaluations of its performance in an additive noise scenario. The problem of optimal choice of the parameters of the proposed technique is also explored, and solutions are presented. Though the cover object is assumed to be an image for purposes of illustration, the proposed method is equally applicable for other types of multimedia data, like video, speech or music.

IP telephony has seen a tremendous surge in interest in the last couple of years. Several equipment vendors, application developers and service provides are entering into this emerging market. Regional as well as international standardization organizations have been involved in developing standards in the area, while several other forums have been actively promoting the industry as a whole. However, until recently, mobility and wireless issues have not been considered in detail with the scope of IP telephony, and the focus has been on fixed IP telephony systems. In this paper, we discuss some of the issues that need to be considered when mobility and wireless transport is introduced within IP telephony.

The recent interest in IP telephony has led to a tremendous increase of standardization activities in the area. The three main standards bodies in the area of IP telephony are the International Telecommunication Union's (ITU-T) Study Group (SG) 16, the Internet Engineering Task Force (IETF) and the European Telecommunication Standards Institute's (ETSI) TIPHON project. In addition, forums such as the International Multimedia Teleconferencing Consortium (IMTC), the Intelligent Network Forum (INF), the International Softswitch Consortium (ISC), the Electronic Computer Telephony Forum (ECTF), and the MIT's Internet Telephony Consortium (ITC) are looking into various other aspects that aim at the growth of this industry. This paper describes the main tasks (completed and in progress) undertaken by these organizations. In describing such work, an overview of the underlying technology is also provided.

Subjective evaluations of IP telephony were collected while network performance was varied using an IP network emulator. Packet loss was set at 0%, 1%, or 8%. Mean one-way delay was set at 8 ms, 46 ms, or 156 ms (in addition to system delay of 90 ms). Delay variation (standard deviation) was set to 1 ms, 8 ms, or 17 ms. For combinations of these network settings, a standard mean opinion score (MOS) of speech quality was collected from 31 human judges. MOS is reported as a function of the network emulator settings.

Originally designed for data communication, the Internet is being considered for real-time voice and video communication (in applications such as telephony, voice conferencing, videophone and video conferencing). In order to become a legitimate alternative to the traditional circuit-switched telephone network, the internet must guarantee toll quality for voice communication, which translates into stringent requirements on voice packets, in particular in terms of delay, loss rate and service availability. In this paper, we seek to assess voice delay in the Internet. In this respect, we use network simulation to derive the end-to-end delay incurred by voice traffic for various network scenarios (i.e. path lengths, link bandwidths) and loading conditions, using different encoding and packetization schemes. As a consequence of the tradeoffs revealed by the delay results obtained, we propose a simple mechanism for end-to-end bandwidth management and admission control for voice streams.

Two standards currently compete for the dominance of IP telephony signaling: the H.323 protocol suite by ITU-T, and the Session Initiation Protocol (SIP) by IETF. Both of these signaling protocols provide mechanisms for call establishment and teardown, call control and supplementary services, and capability exchange. We investigate and compare these two protocols in terms of Functionality, Quality of Service (QoS), Scalability, Flexibility, Interoperability, and Ease of Implementation. For fairness of comparison, we consider similar scenarios for both protocols. In particular, we focus on scenarios that involve a gatekeeper for H.323, and a Proxy/Redirect server for SIP. The reason is that medium-to-large IP Telephony systems are not manageable without a gatekeeper or proxy server. We consider all three versions of H.323. In terms of functionality and services that can be supported, H.323 version 2 and SIP are very similar. However, supplementary services in H.323 are more rigorously defined, and therefore fewer interoperability issues are expected among its implementations. Furthermore, H.323 has taken more steps to ensure compatibility among its different versions, and to interoperate with PSTN. The two protocols are comparable in their QoS support [similar call setup delays, no support for resource reservation or class of service (CoS) setting], but H.323 version 3 will allow signaling of the requested CoS. SIP's primary advantages are (1) flexibility to add new features, and (2) relative ease of implementation and debugging. Finally, we note that H.323 and SIP are improving themselves by learning from each other, and the differences between them are diminishing with each new version.

Conference control is an integral part in many-to-many communications that is used to manage and co-ordinate multiple users in conferences. There are different types of conferences which require different types of control. Some of the features of conference control may be user invoked while others are for internal management of a conference. In recent years, ITU (International Telecommunication Union) and IETF (Internet Engineering Task Force) have standardized two main models of conferencing, each system providing a set of conference control functionalities that are not easily provided in the other one. This paper analyzes the main activities appropriate for different types of conferences and presents an architecture for conference control called GCCP (Generic Conference Control Protocol). GCCP interworks different types of conferencing and provides a set of conference control functions that can be invoked by users directly. As an example of interworking, interoperation of IETF's SIP and ITU's H.323 call control functions have been examined here. This paper shows that a careful analysis of a conferencing architecture can provide a set of control functions essential for any group communication model that can be extensible if needed.

Leaders within the Information Technology (IT) industry are expressing a general concern that the products used to deliver and manage today's communications network capabilities require far too much effort to learn and to use, even by highly skilled and increasingly scarce support personnel. The usability of network management systems must be significantly improved if they are to deliver the performance and quality of service needed to meet the ever-increasing demand for new Internet-based information and services. Fortunately, recent advances in spoken language (SL) interface technologies show promise for significantly improving the usability of most interactive IT applications, including network management systems. The emerging SL interfaces will allow users to communicate with IT applications through words and phases -- our most familiar form of everyday communication. Recent advancements in SL technologies have resulted in new commercial products that are being operationally deployed at an increasing rate. The present paper describes a project aimed at the application of new SL interface technology for improving the usability of an advanced network management system. It describes several SL interface features that are being incorporated within an existing system with a modern graphical user interface (GUI), including 3-D visualization of network topology and network performance data. The rationale for using these SL interface features to augment existing user interfaces is presented, along with selected task scenarios to provide insight into how a SL interface will simplify the operator's task and enhance overall system usability.

This paper describes advanced latency reduction architecture to serve MPEG-1 layer 3 (MP3) sound stream over low bit rate links with IPv6 PPP and its experimentation. Our new technique totally manages late frames and delay jitters in conjunction with rate adjustable stream service in application and jitter control mechanism in data link layer. We have implemented a new IPv6 PPP and rate control scheme experimentally under BSD Unix to fit slow links. We have also measured performance to decide QoS parameters carefully so it will be suitable for use in the real world.

The high data rates required to properly support emerging multimedia internet applications far exceed the capabilities of voiceband modems. For example, seamless real-time delivery of digital video clips might require connections up to 100 times faster. Asymmetric digital subscriber line (ADSL) modems provide residential internet users with a much-needed solution to this remote access bandwidth shortage. ADSL modems connected directly to the internet are installed in the telephone company central office, allowing remote access over the copper twisted pair telephone line by remote transceivers in the homes of residential customers. Because the internet data does not have to pass through the telephone switching network, the ADSL link can provide data rates in the Mbit/sec range. However, for ADSL modems to gain broad acceptance an displace their low rate voiceband counterparts, low cost, ease of installation, and high quality of service must be achieved.

We present an MPEG-2 to H.263 transcoder that accepts an interlaced MPEG-2 bitstream as the input and produces a lower-bitrate progressive H.263 bitstream as the output. As both DVD and digital television may use MPEG-2 interlaced sequences, a potential application of such a transcoder is the transmission of a digital television signal over a wireless medium. Another application is transcoding interlaced DVD content for use on lower-resolution thin clients with progressive displays. The proposed algorithm exploits the properties of the MPEG-2 and H.263 compression standards to perform interlaced to progressive (field to frame) conversion with spatial downsampling and frame-rate reduction in a CPU and memory efficient manner, while additionally minimizing picture quality degradation as measured by PSNR. This is the first algorithm to our knowledge that effectively uses both spatial and temporal downsampling in an MPEG-2 to H.263 field to frame transcoder in order to achieve substantial bitrate reduction. This paper discusses recoding experiments used to determine appropriate source and target coding parameters for the transcoder, provides a detailed description of the transcoding algorithm, and describes the performance of a software implementation of the transcoder.

It is tempting to predict that the Internet will never be able to support the universal delivery of interactive, high quality, long format video to serve the majority of the population. After all, it would severely stress the capability of the Internet. Further, we enjoy well established television, cable, and recorded video industries that deliver a variety of program material for entertainment, information, and education. These industries continue to evolve to deliver ever-increasing convenience and variety of material. On the other hand, especially given the momentum of the last few years, it is equally tempting to predict that the Internet will subsume all forms video delivery. How do we know where we are on this range of alternatives?

Video compression reduces the required bit rate and thereby enables a number of applications such as digital television and video over the Internet. However, compression also makes the video much more susceptible to errors, e.g. bit errors or packet loss. Generally there is a tradeoff between the compression of a signal and the error resilience of the compressed representation. However, recently there have been a number of developments which provide significant improvements in error-resilience while resulting in relatively small reductions in compression. This paper provides an overview of current error-resilient compression methods and also proposes an error-resilient coding scheme referred to as Multiple State Streams (MSS). Conventional video compression standards employ a similar architecture which we refer to as single-state systems since they have a single state (e.g. the previously decoded frame) which if lost or corrupted adversely affects all subsequent frames until the state is reinitialized (the prediction is refreshed). We propose to combat this problem of incorrect state at the decoder with MSS by coding the video into multiple independently decodable streams, each with its own prediction process and state, such that if one stream is lost the other streams can still be used to produce usable video. More importantly, the correctly received streams provide improved error concealment and enable faster state recovery for the lost stream.

Current Video on Demand (VoD) systems do not provide for interactive VCR-like scan operations such as rewind or fast- forward. Pre-storing the entire movie on a set-top box for implementation of such operations is currently neither practical nor feasible. Other approaches to VCR-like controls are either computationally complex or the server or consume large amounts of bandwidth during the operation. This paper proposes a novel technique, VVD, for real-time implementation of VCR-like operations, supporting multiple scan speeds, for Video on Demand with low latency and overhead. VVD uses multi-level compression of movie files to encode the same movie at a hierarchy of playout rates. Depending on the rate at which a scan operation is requested, the corresponding compressed movie file is transmitted. Frame dropping and repetition are used to provide smooth speed transitions among the multiple levels. The overhead of this scheme can be controlled to be less than 100% of the movie storage size.

Motion estimation constitutes a critical part in video coding since it is computationally expensive and has a substantial impact on the resulting image quality. A very efficient method to estimate motion is through adding a rate constraint for coding of the motion vectors into the estimation process. In this paper we analyze the performance of this approach in the emerging MPEG-4 standard, considering frame- as well as object-based video sequences. The investigations show that rate constrained motion estimation generally offers significant gains in coding efficiency which, however, depend on the shape of the object coded. Complexity measurements further show that a rate constraint imposes only a marginal computational overhead, thus making this approach a useful amendment for an MPEG-4 coder.

The recently developed MPEG-4 technology enables the coding and transmission of natural and synthetic audio-visual data in the form of objects. In an effort to extend the object-based functionality of MPEG-4 to real-time Internet applications, architectural prototypes of multiplex layer and transport layer tailored for transmission of MPEG-4 data over IP are under debate among Internet Engineering Task Force (IETF), and MPEG-4 systems Ad Hoc group. In this paper, we present an architecture for interactive MPEG-4 speech/audio transmission system over the Internet. It utilities a framework of Real Time Streaming Protocol (RTSP) over Real-time Transport Protocol (RTP) to provide controlled, on-demand delivery of real time speech/audio data. Based on a client-server model, a couple of low bit-rate bit streams (real-time speech/audio, pre- encoded speech/audio) are multiplexed and transmitted via a single RTP channel to the receiver. The MPEG-4 Scene Description (SD) and Object Descriptor (OD) bit streams are securely sent through the RTSP control channel. Upon receiving, an initial MPEG-4 audio- visual scene is constructed after de-multiplexing, decoding of bit streams, and scene composition. A receiver is allowed to manipulate the initial audio-visual scene presentation locally, or interactively arrange scene changes by sending requests to the server. A server may also choose to update the client with new streams and list of contents for user selection.

An error resilient coding method for 3D graphic models is proposed in this research, which exploits the topology and geometry information of the original mesh to convert the mesh to a pre- defined 3D structure via morphing, and then partitions the morphed structure into a set of smaller pieces via volume splitting. By using the morphing and splitting processes, the error propagation length is greatly reduced since it is confined to each individual piece rather than the whole mesh. Different pieces are connected via their joint boundaries. By coding the joint boundaries separately, individual pieces can be reconstructed independently of each other. Consequently, error resilient coding can be conveniently applied to each piece. The number of pieces of partitioning is determined by channel error rates. It is demonstrated that the proposed method can achieve satisfactory results in an environment of bit error rates less than 10^-3.

Two studies examined the effect of B-frames on subjective quality of MPEG-2 video. One study used consumer judgments in a variant of the standard CCIR Recommendation 500-5 procedure for collecting subjective evaluations. The other study used the judgments of a single expert in adjusting the bit rate necessary for MPEG-2 without B-frames to be subjectively equal to MPEG-2 with B-frames at a given bit rate. The results of the two studies were qualitatively similar. Summary of results: Picture quality improved with increase in bit rate until a saturation point was reached. The introduction of B-frames improved picture quality, especially for difficult source material. This was more noticeable at the lower bit rates (e.g., 3 Mb/s). The difference in bit rate between MPEG-2 with and without B-frames varied substantially with the source material. For example, for basketball, 5 Mb/s with B-frames was subjectively equal to 8 Mb/s without B-frames, but for other material the difference was near zero.

The objective measurement of visual quality of DTV imagery may be performed in two basic ways: 'double-ended' and 'single-ended' methods. A 'single-ended' method is generally used to assess the visual quality of a Digital TV signal when the original signal is not available (e.g. a studio receiving many contributions on digital links). This paper proposes a new 'single-ended' system to objectively assess the visual quality of a DTV based on a method for which new technique of detection of the impairments and a temporal analysis of the quality level have been applied. This method tries to go beyond the limit of a bare 8 X 8 block structure and to look for the visibility of an impairment (here called 'snow flakes') due to a temporal fluctuation of visual quality, typical of the MPEG GOP structure. Experimental results show a good behavior when low bit rate (1 Mbps to 4 Mbps) coded video signal are processed. Other experimental results at higher bit rate (5 Mbps to 9 Mbps) are also proposed, for which the results are more dependent on the image content.

The investigation examines two methodologies by which to control the impairment level of digital video test materials. Such continuous fine-tuning of video impairments is required for psychophysical measurements of human visual sensitivity to picture impairments induced by MPEG-2 compression. Because the visual sensitivity data will be used to calibrate objective and subjective video quality models and scales, the stimuli must contain realistic representations of actual encoder-induced video impairments. That is, both the visual and objective spatio-temporal response to the stimuli must be similar to the response to impairments induced directly by an encoder. The first method builds a regression model of the Peak Signal-To-Noise Ratio (PSNR) of the output sequence as a function of the bit rate specification used to encode a given video clip. The experiments find that for any source sequence, a polynomial function can be defined by which to predict the encoder bit rate that will yield a sequence having any targeted PSNR level. In a second method, MPEG-2-processed sequences are linearly combined with their unprocessed video sources. Linear regression is used to relate PSNR to the weighting factors used in combining the source and processed sequences. Then the 'synthetically' adjusted impairments are compared to those created via an encoder. Visual comparison is made between corresponding I-, B-, and P-frames of the synthetically generated sequences and those processed by the codec. Also, PSNR comparisons are made between various combinations of source sequence, the MPEG-2 sequence used for mixing, the mixed sequence, and the codec-processed sequence. Both methods are found to support precision adjustment of impairment level adequate for visual threshold measurement. The authors caution that some realism may be lost when using the weighted summation method with highly compression-impaired video.

Many organizations have focused on developing digital video quality metrics which produce results that accurately emulate subjective responses. However, to be widely applicable a metric must also work over a wide range of quality, and be useful for in-service quality monitoring. The Institute for Telecommunication Sciences (ITS) has developed spatial-temporal distortion metrics that meet all of these requirements. These objective metrics are described in detail and have a number of interesting properties, including utilization of (1) spatial activity filters which emphasize long edges on the order of 10 arc min while simultaneously performing large amounts of noise suppression, (2) the angular direction of the spatial gradient, (3) spatial-temporal compression factors of at least 384:1 (spatial compression of at least 64:1 and temporal compression of at least 6:1, and 4) simple perceptibility thresholds and spatial-temporal masking functions. Results are presented that compare the objective metric values with mean opinion scores from a wide range of subjective data bases spanning many different scenes, systems, bit-rates, and applications.

The effectiveness of statistical multiplexing was investigated in two ways. One the one hand, two-channel tests facilitated detailed statements about the mechanisms of statistical multiplexing. Based on a simple concept of criticality for MPEG-2 coding, good agreement with theoretical expectations was found only for combinations of sequences which differed strongly in criticality. For these cases, a considerable increase in quality of the more critical sequence sand even an increase of the 'overall quality' (quality of sequences of both channels) was observed. On the other hand, the improvement of the picture quality was evaluated for multi-channel systems (4 and 8 channels). In a life demonstration at an international assembly, many experts could convince themselves of the advantage of this method. The results were also recorded and evaluated using a recently developed method of objective picture quality measurement which confirmed the subjective impression and led to quantitative statements. It became quite evident that peaks of impairment could be avoided by statistical multiplexing.

End-to-end quality of service for real-time video delivery over ATM networks depends on both video coding and network performance. Historical measures of QoS for data services have involved packet delivery parameters such as cell loss and delay, but these are not directly representative of a user's perception of the delivered video quality. To this end, this paper describes the use of a video quality measurement scheme that employs a human vision based picture quality model to objectively estimate end-user video quality. We present results to demonstrate that the video quality analysis based approach can be used to determine performance characteristics for distribution-quality video coding parameters over ATM networks.

This paper presents a method designed to monitor image quality. The emphasis is given here to the monitoring in digital television broadcasting networks, in order for the providers to ensure a 'user-oriented' Quality of Service. Most objective image quality assessment methods are technically very difficult to apply in this context because of bandwidth limitations. We propose a parametric, reduced reference method that relies on the evaluation of characteristic coding and transmission impairments with a set of features. We show that quality can be predicted with a satisfying correlation to a subjective evaluation by the combination of several impairment features in an appropriate model. The method has been implemented and tested in a range of situations on simulated and real DVB networks. This allows to conclude on the usefulness of the approach and our future developments for quality of service monitoring in digital television.

The current Internet is not well suited for the transmission of high quality video such as MPEG-2 because of severe quality degradation during network congestion episodes. One possible solution is the combination of layered video coding with the Differentiated Services (DiffServ) architecture; different video layers are mapped into different priority levels, and packets with different priorities receive a different dropping treatment in the network. It is expected that with layering and priority dropping, graceful degradation of video quality will be experienced during congestion episodes. We consider various layering mechanisms defined in the MPEG-2 standards; namely, temporal scalability, data partitioning (DP) and Signal to Noise Ratio (SNR) scalability. The main issue in this paper is how layers should be created to maximize perceived video quality over a given range of network conditions. Key to our study is the use of real life video sequences and a video quality measure consisting of a perceptual distortion metric based on the Human Visual System (VHS). Our results show that video quality is sensitive to how layering is accomplished, and that there is an optimum layering that maximizes the quality for a given network condition. Our results also show that layering can achieve higher network loading for a given minimum quality target than non-layered video, and can achieve graceful degradation over a wider range of network conditions. We have also seen that the wider the range of network conditions is, the higher is the number of layers required in order to remain at the highest possible quality level for each network condition. In particular, we demonstrate how three or four layers achieve better results than two layers; however, additional layers beyond four provide marginal improvement. Therefore, from a practical point of view, three or four layers are sufficient to attain most of the benefits of layering. We compare the various scalability techniques in terms of complexity and video quality. Temporal scalability, which restricts the layering to be done at frame boundaries, is the simplest to implement and introduces no overhead, but performs poorly compared to data partitioning, which allows the grouping of coefficients into layers independent of the frames they belong to. This shows that, contrary to customary belief, dropping data in B frames prior to dropping data in P or I frames is a poor layering technique. DP is much simpler to implement and introduces significantly lower overhead than SNR scalability. However, SNR scalability provides higher quality than DP when network conditions are particularly poor.

In distributed video applications, proper transmission scheduling is necessary to ensure display continuity and efficient bandwidth utilization. Many optimization algorithms exist for the transmission planning of stored VBR videos. Most of these algorithms operate under the assumption of jitter free transmissions. In practical applications, however, such an assumption may not always be true due to specific server and network conditions such as server load shifting and mobile connection handoffs. Under such conditions, video transmission would enter temporary inertia periods in which no data can be transmitted. To cope with such situations, we propose in this paper transmission planning algorithms which ensure uninterrupted client playback and smooth network traffic. We consider different jitter scenarios in practical video applications and formulate optimization problems based on these scenarios. Optimal or near optimal solutions are then presented for these problems and the effectiveness of these algorithms are tested using available VBR video traces.

Rapid advantages in computer and telecommunication technologies have made integrated services packet-switched networks possible. It is expected that a significant portion of future networks will carry prerecorded video. Traffic in applications such as Video on Demand (VoD), will include high-fidelity audio, short multimedia clips, and full-length movies. In this paper we study the effect of video rate smoothing of a prerecorded VBR sources, on network utilization and statistical multiplexing gain, such that the end users receive satisfactory service. The enhancement piecewise constant rate transmission and transport (e-PCRTT) algorithm is used as the smoothing algorithm which utilizes equal size intervals and its effect on the statistical characteristics of the multiplexed video is explored. We found that synchronization between the smoothing intervals can significantly improve the efficiency of the smoothing process by reducing the number of bandwidth changes and the rate variability of the multiplexed stream. We investigate how synchronized smoothing intervals influence the potential for statistical gain, as compared to unsynchronized streams and present several examples to illustrate the advantage of synchronized streams over unsynchronized streams.

Multiservice Internet Protocol-based Virtual Private Networks (MIP- VPNs) with Quality of Service (QoS) are becoming a reality due to the availability of new standards from the Internet Engineering Task Force (IETF). This paper describes how components including security models, IP tunneling protocols, and service differentiation schemes fit together in order to construct such a VPN. First, the concept and rationale of VPN is presented, followed by a discussion of its supporting components. A comparison is made among the various VPN technologies.

We address a new error-resilient scheme for broadcast-quality interactive MPEG-2 video streams to be transmitted over lossy packet networks. A new scene-complexity adaptive mechanism, namely Adaptive MPEG-2 Information Structuring and Protection (AMISP) is introduced. AMISP lies on an information structuring scheme which modulates the number of resynchronization points (i.e., slice headers and intra-coded macroblocks) in order to maximize the perceived video quality. The video quality the end-user experiences depends both on the quality of the compressed video before transmission and on the degradation due to packet loss. Therefore, the structuring scheme constantly determines the best compromise between the rate allocated to encoding pure video information and the rate aiming at reducing the sensitivity to packet loss. It is then extended with a Forward Error Correction (FEC) based protection algorithm to become AMISP. AMISP triggers the insertion of FEC packets in the MPEG-2 video packet stream. Finally, it is shown that AMISP outperforms usual MPEG-2 transmission schemes, and offers an acceptable video quality even at loss ratios as high as 10^-2. Video quality is estimated using the Moving Picture Quality Metric, which proved to behave consistently with human judgment.

Video is inherently frame-oriented and most applications such as commercial video processing require to manipulate video in terms of frames. However, typical video servers treat videos as byte streams and perform random access based on approximate byte offsets to be supplied by the client. They do not provide frame or timecode oriented API which is essential for many applications. This paper describes a frame-oriented architecture for video servers. It also describes the implementation in the context of IBM's VideoCharger server. The later part of the paper describes an application that uses the frame architecture and provides fast and slow-motion scanning capabilities to the server.

The exponential growth and dramatic increase in demand for network bandwidth is expanding the market for broadband satellite networks. It is critical to rapidly deliver ubiquitous satellite communication networks that are differentiated by lower cost and increased Quality of Service (QoS). There is a need to develop new network architectures, control and management systems to meet the future commercial and military traffic requirements, services and applications. The next generation communication networks must support legacy and emerging network traffic while providing user negotiated levels of QoS. Network resources control algorithms must be designed to provide the guaranteed performance levels for voice, video and data having different service requirements. To evaluate network architectures and performance, it is essential to understand the network traffic characteristics.

Robust and adaptive operation of a video conferencing system can be realized by proper utilization of a network of video cameras and microphone arrays. Such multimodal sensory modules can provide valuable information about the geometrical and acoustical properties of an environment as well as can allow for real-time monitoring for dynamic activity in the environment. In this paper, we present an overview of the research activities focused on utilization of omnidirectional camera networks for geometrical and environmental modeling and microphone arrays for speaker localization.

The communication and cooperation of widely distributed applications is enabled by conference control services for tightly- coupled environments like SCCS (Scalable Conferencing Control Service) providing a tight control of resources for application state synchronization and multipoint communication in the conference. A major issue in the design of such services is the scalability in terms of users and distribution. Especially the used resource management scheme is crucial for a high scalability. SCCS uses an own-developed scheme which improves the response time of resource requests especially if they are local in the topology. For further improvement of that scheme the reconfiguration of existing topologies is another issue. The problem of performing a reconfiguration (without blocking the conference) is not within the scope of this paper. The scope is the need for a dynamic decision how and if the conference should be reconfigured. For that a scheme is presented which detects active entities in the conference and groups them closer to each other in the topology to improve further requests. A definition of 'activity' is given together with a scheme to identify that activity during a running conference. For that, unique sets of active entities in the running conference are determined built according to the used resources in the sets. We present a placement algorithm for those active entity sets to optimize further resource requests for the specific resources used by the corresponding entity set. A scheme is presented which is the basis for the decision to perform a reconfiguration and it is embedded in the conferencing service SCCS. Additionally a performance evaluation is presented using an automata-based load model to determine the performance gain in different scenarios.

The present paper addresses the main features and performance test results of the CONCHA -- CONference system based on java and corba event service CHAnnels, developed at the Communications and Telematic Services Group of the University of Coimbra. The conference framework was conceived using the Java^TM platform, based on the ITU T.120 recommendations, with reliable multicasting communication for both data and control information. The conference control service extends the CORBA Event Service architecture with mechanisms for native IP multicasting mapping, optimization of reliability mechanisms and processes, filtering, total ordering of events/messages and fragmenting/reassembling of large messages.

The steadily increasing quality of hard- and software enables the development of more and more sophisticated groupware. However, modern distributed systems and distributed multimedia systems in particular challenge the abilities of groupware manufacturers, due to the complexity involved in the development of those systems. Frameworks for collaborative environments might provide a solution to these problems, helping to reduce complexity in groupware development by provision of suitable software components. This article will identify the demands and requirements of distributed (multimedia) systems and their human users that frameworks have to take into account. The article then introduces the HOUCOM framework for collaborative environments and sketches the employment of this framework in a sample scenario.

The intra-media synchronization control scheme for MPEG video over Internet, where communication quality is not always guaranteed, are proposed in this paper. In order to obtain expected quality of MPEG video, four basic control functions and one subfunction for a minimum size of controlled image (hereinafter referred to as MU: Media Unit) are introduced. The proposed basic control functions are classified into two types. One type lays emphasis on realizing real-time display. Another type lays emphasis on keeping quality of images. The former type includes short (to reduce periods to display MUs), change (to realize more moderate control than short function) and skip (to abandon a received MU) functions. The latter type includes smooth (to make the variations of periods to display MUs small by postponing a display time) function. And a sort function, which is to display MUs in regular order, is also prepared in the proposed control scheme. Combinations of these control functions are pursued at a client side alone. Characteristics of proposed control functions are described with simulation results. Especially, in order to execute the subjective assessment only depending on psychological assessment factors, a SD (Semantic Differential) method, which is a kind of psychological factor analysis, is introduced. As the result, it is confirmed that a smooth function is the most effective to keep quality of images, and in order to realize realtime display, the short, change or skip function should be combined.

In this paper we present a new method for indexing video data based on the descriptions of object behaviors within a scene. The behaviors of objects over time is the most important feature in a scene and make up the semantic contents of the scene. The indices generated based on these temporal features are complementary to still image features for achieving fast and efficient search and retrieval of the video contents. To provide the conceptual framework for indexing based on the object behaviors we propose an object Behavior Description Scheme (BDS) for video data. The Behavior Description Scheme is based on spatio-temporal features and the associated object behavior descriptions are directly extractable from the video data itself (and support data) facilitating automatic indexing.

In this paper, we describe Java Virtual Collaboration Development Environment (JVCDE) for building synchronous collaborative application in heterogeneous environments. The meaning of collaboration in this paper is an application sharing mechanism that allows multiple users to simultaneously work in an existing single user application. In collaboration system, one of the major problems is to support interoperability in multiple environments, where, all the participants of collaboration can share the same results. But In heterogeneous environments, it is more difficult to guarantee interoperability than homogenous environments. To solve these interoperability problems, we design and implement Java Virtual Collaboration Development Environment using JAVA language. We develop a prototype system based on the proposed framework.

One major problem of using multimedia material in lecturing is the trade-off between actuality of the content and quality of the presentations. A frequent need for content refreshment exists, but high quality presentations can not be authored by the individual teacher alone at the required rate. Several past and current projects have had the goal of developing so-called learning archives, a variation of digital libraries. On demand, these deliver material with limited structure to students. For lecturing, these systems provide just as insufficient service as the unreliable WWW. Based on our system HyNoDe [HYN97] we address these issues in our distributed media server built of 'medianodes.' We add content management that addresses teachers' needs and provide guaranteed service for connected as well as disconnected operation of their presentation systems. Medianode aims at a scenario for non-real-time, shared creation and modification of presentations and presentation elements. It provides user authentication, administrative roles and authorization mechanisms. It requires an understanding of consistency, versioning and alternative content tailored to lecturing. To allow for predictable presentation quality, medianode provides application level QoS supporting alternative media and alternative presentations. Viable presentation tracks are dynamically generated based on user requests, user profiles and hardware profiles. For machines that are removed from the system according to a schedule, the systems guarantees availability of consistent, complete tracks of selected presentations at disconnect time. In this paper we present the scope of the medianode project and afterwards its architecture, following the realization steps.

In this note we demonstrate the utility of omni-directional images for efficiently deriving high-resolution environment models. Omni- Directional Vision Sensors (ODVS), which have a wide field of view, are the most suitable for immersive virtual walk-through in a remote environment. In this paper, we show omni-directional images that are acquired using high-resolution digital camera. Also we show the synthetic omni-directional images generated using 3D graphics methods. The high-resolution and the synthetic images provide invaluable high-quality input image for visual modeling research use.

In digital video compression, motion estimation is most commonly used in reducing temporal redundancy. The motion information generated by motion estimation is usually coded in the compressed stream. In many situations, the compressed video needs to be post- processed, for example, filtered to generate an out-of-focus effect. The filtered video then requires another motion estimation to be coded in bit stream. We propose an algorithm to reuse the existing motion information in the original video to avoid or aid the motion estimation in the recompression. Since this technique significantly reduces post-processing time for compressed video when the output is required in compressed form, it is highly desirable in digital video broadcasting systems.