This work presented a novel vision-based augmented-reality system for applications in multiuser collaborative environments. The kernel technology of this vision-based system locates the cameras that are utilized to point and simulate the positions of multiple viewers. Camera calibration based on computer vision is employed during the camera's locating process. The applications in multiuser collaborative environments allow the viewers to view various positions and in numerous directions. However, traditional calibration approaches are not suitable for these cases sufficiently. A novel calibration pattern based on pseudo-random arrays is designed for multiuser collaborative applications. The pattern has a simple and regular structure, easily extracts features, achieves robust recognition using local information, and does not limit viewer positions and directions. Experimental results indicate that the proposed system provides a effective platform for applications in multiuser collaborative environments.

Proxy-caching strategies, especially prefix caching and interval caching, are commonly used in video-on-demand (VOD) systems to improve both the system performance and the playback experience of users. However, because these caching strategies are designed for homogeneous clients, they do not perform well in the real world where clients are heterogeneous (i.e., different available network bandwidths and different sizes of client-side buffers). This paper investigates the problems caused by heterogeneous client-side buffers. We analyze the theoretical performance of these caching strategies, and then, derive cost functions to measure the corresponding performance gains. Based on these analytical results, we develop a caching strategy that employs both prefix caching and interval caching to minimize the input bandwidth of a proxy. The simulation results demonstrate that the bandwidth requirements of a proxy implementing our caching strategy are significantly lower compared to adopting prefix caching or interval caching alone.

Mode decision is a kernel function of current video standards, especially while dealing with massive coding modes as the latest H.264/AVC variable block size partitions. The mode decision process is extreme time consuming because it contains the kernel motion estimation and intra prediction modules. By applying the concepts of using priority coding orders to encode a sequence, this paper concludes a set of mode decision strategies with observing among priority lists. A computation-aware mode decision scheme is then proposed to the H.264/AVC. As the results demonstrated, coding with priority orders can achieve computation-aware ability automatically and the priority info is very helpful for developing software-based computation-aware mode decision algorithms.

Content with rapid, moderate, and slow motion is frequently mixed together in real video sequences. Until now, no fast block-matching algorithm (FBMA), including the well-known three-step search (TSS), the block-based gradient descent search (BBGDS), and the diamond search (DS), can efficiently remove the temporal redundancy of sequences with wide range motion content. This paper proposes an adaptive FBMA, called A-TDB, to solve this problem. Based on the characteristics of a proposed predicted profit list, the A-TDB can adaptively switch search patterns among the TSS, DS, and BBGDS, according to the motion content. Experimental results reveal that the A-TDB successfully adopts the search patterns to remove the temporal redundancy of sequences with slow, moderate and rapid motion content.

By forwarding the server stream client by client, a chaining-based scheme is a good way to reduce the server streams for streaming applications in well-connected networks. In this paper, we prove that the minimum number of required server streams in such schemes is n-k+1, where n is the number of client requests and k is a value determined by client buffer sizes and the distribution of requests. In addition, we present an optimal chaining algorithm using a dynamic buffer allocation strategy. Compared to existing chaining schemes, our scheme not only utilizes the backward (basic chaining) and/or forward (adaptive chaining) buffer, but also exploits the buffers of other clients in order to extend the chain as much as possible. In this way, more clients can be chained together and served by the same server stream. Our simulation results show that the requirements of the server streams in the presented scheme are much lower those of existing chaining schemes. We also introduce mechanisms for handling VCR functions and fault exceptions in practical applications.

The fast block-matching algorithm (FBMA), includes the well-known three-step search (TSS), the block-based gradient descent search (BBGDS) and the diamond search (DS), have been developed for motion estimation in designing a video codec. In this paper, an adaptive FBMA, named A-TDB, to solve the block matching problem for varying motions is proposed. Based on the characteristics of a proposed predicted profit list, the A-TDB can adaptively switch the search patterns among the TSS, DS, and BBGDS according to the motion contents.

On the basis of the emerging peer-to-peer structure and the characteristics of user behavior, this paper proposes a scalable video delivery scheme, named P2PVD, for large scale video-on-demand application. In P2PVD, two types of orders are permitted: reserved and urgent orders. Reserved orders are encouraged with a lower price policy to smooth the network traffic and the server load. A delay-aware dynamic parallel transmission in the requesting peer is introduced to serve the reserved and urgent orders. The supplying peer is carefully designed based on three priority rules to increase the capacity of P2PVD. Experimental results indicate that P2PVD is scalable.

A simple and efficient motion-vector-based search engine supporting query by clip is presented for indexing and retrieving of videos. We propose a technique to filter and reconstruct noisy motion vectors to improve the fidelity of the motion activity. In addition to the motion features of intensity and direction, a new motion descriptor, named representative spectrum, extracted by a vector quantization expressing the spatial-temporal distribution of motion activity, is introduced for video indexing. A dynamic programming scheme is also utilized to measure the similarity of two videos based on the concept of common subsequence. Furthermore, a 2-step search of similar videos is employed to reduce the computational complexity. The experimental results indicate that the proposed search engine performs well and the obtained recall and precision values are also high enough.

In this paper, a new dual-bitstream technique for video cassette recording (VCR) functionality and a transcoding technique of the motion vectors (mv) between dual bitstreams are introduced. In the dual-bitstream technique, a direct reference bitstream is proposed to replace the original reverse bitstream to reduce the required network bandwidth and decoder effort of VCR requests. In the transcoding technique, the mv in the forward bitstream are first reused to generate the mv in the reverse bitstream and the direct reference bitstream. If the obtained error is larger than a preset threshold, the mv are refined step-by-step from the inner to the outer of a refinement window.

The performance of streaming caches is not only affected by cache management but also determined by cache cooperation. If caches cooperate together, the desired video may be obtained from other caches thus reducing server load. However, such process increases network traffic between caches. It must be done carefully. This paper proposes a cooperative cache framework to automatically select appropriate cache sources and paths for various network environments. Simulation results demonstrate that the proposed framework is efficient and easy to implement for streaming caches.

Many fast block-matching algorithms (BMAs) reduce the computational complexity of motion estimation by sophisticatedly inspecting a subset of checking points, and stop only when all those checking points have been examined. This means that the searching process for each current block cannot be interrupted, even when it is performed in a software-based computation environment. Our main goal is to allow the searching process to stop once a specified amount of computation has been performed. A novel computation-aware scheme is proposed, which first dynamically determines the target amount of computation power allocated to a frame, and then allocates this to each block in a computation-distortion-optimized manner. We propose a rate-control-like procedure and a predicted computation-distortion benefit heuristic to realize this scheme. Conventional BMAs, such as full-search block matching, three-step search, new three-step search, four-step search, and diamond search, can be transformed into their corresponding computation-aware BMA versions. In our simulations, the resulting computation-aware BMAs not only exhibit higher efficiency than conventional BMAs, but also allow the motion estimation to terminate after any specified amount of computation has been performed (in units of checking points).

Chaining is a promising technique in alleviating the network and the server load by serving numbers of clients in a single server stream. In this scheme, each client temporarily stores the ongoing media stream in the client buffer and forwards it to another client with adequate buffering delay to meet the next client playback time. Therefore, the clients request the same video program can be chained together if the client buffer is large enough to bridge clients' playback time. However, existing chaining schemes do not efficiently exploit the client buffer space. In this paper, we present an optimal chaining algorithm to fully utilize the client buffer space and hence minimize the required server streams. Our performance comparison demonstrates that the optimal chaining scheme significantly reduces the server load over the existing chaining schemes. More importantly, the proposed scheme can integrate the streaming proxy into the video chain, which makes the video chain more reliable and flexible in real systems. This paper also presents the prototype implementation of the optimal chaining based video-on-demand system.

Video-on-Demand (VoD) systems are attractive multimedia services over the Internet. However, conventional VoD systems can not serve many viewers simultaneously because such large-scale services require considerable server load and network bandwidth. This study proposes a cooperative proxy scheme, consisting of a streaming cache and a query protocol revised from Internet Cache Protocol (ICP), to improve performance of large-scale VoD systems. The streaming cache we proposed attempts to serve the most requests with slightly computation overhead In addition, our query protocol further extends service scalability by satisfying requests "as far as possible", although Web proxies prefer "as near as possible". Simulation results demonstrate that our proxy scheme is efficient especially in large-scale VoD systems.

This paper presents an integrated systolic array design for implementing full-search block matching, 2-D discrete wavelet transform, and full-search vector quantization on the same VLSI architecture. The above three functions are prime essential but with a large amount of computation in video compression. To meet real-time requirements, many systolic arrays are designed for individually performing each of them. In fact, these functions contain the similar computational procedure. The matrix-vector product forms of them are quite similar. In this paper, by carefully extracting the common computation component, an integrated systolic array design that can perform the above three functions is presented. A utilization of 100% to 97% is achieved for executing full-search block matching and full-search vector quantization. When performing the 2-D discrete wavelet transform, the utilization is about 32%. The proposed integrated architecture has lower hardware cost and has a a regular hardware structure. It befits the VLSI implementation for video compression.

In this paper, we propose a novel fast block matching algorithm design strategy based on the complexity-distortion optimization. The goal of our design is to achieve the maximum image quality under a target computational complexity and to provide a flexible fast block matching algorithm that allows user to terminate algorithm at any target computational complexity. Based on the proposed predictive complexity-distortion benefit list technique, which is employed to predict the motion compensation benefit, we modify the full-search block matching, three-step search, new three-step search, and four-step search to the flexible algorithm. Experimental results show that the flexible algorithms could achieve better performance than traditional fixed complexity algorithms from the viewpoint of complexity-distortion optimization.

In this correspondence, a mean residual vector quantizer (MRVQ) is superimposed on the motion estimator for improving the intraframe coding quality of MPEG. The proposed scheme performs an MRVQ before executing the discrete cosine transformation (DCT) for the purpose of curing the DCT intrinsic trouble of blocking effect at low bit rate.

In recent years, the variable block size (VBS) motion estimation technique has been widely employed to improve the performance of the block matching algorithm (BMA). In VBS, the block size is varied according to the type of motion. The VBS is known to be very efficient for areas containing complex motions. But, it requires a large amount of computation. The computation required is 2 or 5 times compared with the full search BMA. In this paper, a new VBS motion estimation algorithm, called the classified variable block size (CVBS), is proposed to overcome this problem. The algorithm classifies the input blocks into 3 categories: the background, the shade-motion, and the edge-motion. According the characteristics of the classified blocks, various motion estimation techniques are then used to improve the coding performance. The performance of VBS and CVBS under the coding efficiency is investigated. It is shown that the CVBS algorithm requires about 1/6 to 1/10 computations needed by the conventional VBS algorithm.>

Finite state vector quantization (FSVQ) has been proven during recent years to be a high quality and low bit rate coding scheme. A FSVQ has achieved the efficiency of a small codebook (the state codebook) VQ while maintaining the quality of a large codebook (the master codebook) VQ. However, the large master codebook has become a primary limitation of FSVQ if the implementation is carefully taken into account. A large amount of memory would be required in storing the master codebook and also much effort would be spent in maintaining the state codebook if the master codebook became too large. This problem could be partially solved by the mean/residual technique (MRVQ). That is, the block means and the residual vectors would be separately coded. A new hybrid coding scheme called the finite state residual vector quantization (FSRVQ) is proposed in this paper for the sake of utilizing both advantage in FSVQ and MRVQ. The codewords in FSRVQ were designed by removing the block means so as to reduce the codebook size. The block means were predicted by the neighboring blocks to reduce the bit rate. Additionally, the predicted means were added to the residual vectors so that the state codebooks could be generated entirely. The performance of FSRVQ was indicated from the experimental results to be better than that of both ordinary FSVQ and RMVQ uniformly.