In this paper, we propose a new fast algorithm for block motion vector (MV) estimation based on the correlations of the MV's existing in spatially and temporally adjacent as well as hierarchically related blocks. We first establish a basic framework by introducing new algorithms based on spatial correlation and then spatio-temporal correlations before integrating them with a multiresolution scheme for the ultimate algorithm. The main idea is to effectively exploit the information obtained from the corresponding block at a coarser resolution level and spatiotemporal neighboring blocks at the same level in order to select a good set of initial MV candidates and then perform further local search to refine the MV result. We show with experimental results that, in comparison with the full search algorithm, the proposed algorithm achieves a speed-up factor ranging from 150 to 310 with only 2-7% mean square error (MSE) increase and a similar rate-distortion performance when applied to typical test video sequences.

PrefaceThe demand for multimedia wireless communications is growing today at an explosive pace. One common feature of many current wireless standards for high-rate multimedia transmission is the adoption of a multicarrier air interface based on either orthogonal frequency-division multiplexing (OFDM) or orthogonal frequency-division multiple-access (OFDMA). The latest examples of this trend are represented by the IEEE 802.11 and IEEE 802.16 families of standards for wireless local area networks (WLANs) and wireless metropolitan area networks (WMANs). Although the basic principle of OFDM/OFDMA is well established among researchers and communication engineers, its practical implementation is far from being trivial as it requires rather sophisticated signal processing techniques in order to fully achieve the attainable system performance. This book is intended to provide an accessible introduction to OFDMbased systems from a signal processing perspective. The first part provides a concise treatment of some fundamental concepts related to wireless communications and multicarrier systems, whereas the second part offers a comprehensive survey of recent developments on a variety of critical design issues including synchronization techniques, channel estimation methods, adaptive resource allocation and practical schemes for reducing the peakto-average power ratio of the transmitted waveform. The selection and treatment of topics makes this book quite different from other texts in digital communication engineering. In most books devoted to multicarrier transmissions the issue of resource assignment is not discussed at all while synchronization and channel estimation are only superficially addressed. This may give the reader the erroneous impression that these tasks are rather trivial and the system can always operate close to the limiting case of ideal synchronization and channel estimation. However, as discussed Multi-Carrier Techniques for Broadband Wireless Communicationsin this book, special design attentions are required for successfully accomplishing these tasks. In many cases, the trade-off between performance and system complexity has to be carefully taken into consideration in the practical implementation of multicarrier systems. Most of the presented material originates from several projects and research activities conducted by the authors in the field of multicarrier transmissions. In order to keep the book concise, we do not cover advanced topics in multiple-input multiple-output (MIMO) OFDM systems as well as latest results in the field of resource assignment based on game theory. Also, we do not include a description of current wireless standards employing OFDM or OFDMA which are available in many other texts and journal papers.The book is written for graduate students, design engineers in telecommunications industry as well as researchers in academia. Readers are assumed to be familiar with the basic concepts of digital communication theory and to have a working knowledge of Fourier transforms, stochasti...

The fractional Brownian motion (fBm) model has proven to be valuable in modeling many natural processes because of its persistence for large time lags. However, the model is characterized by one single parameter that cannot distinguish between short-and long-term correlation effects. This work investigates the idea of extending selfsimilarity to create a correlation model that generalizes discrete fBm referred to as asymptotic fBm (afBm). Namely, afBm is parameterized by variables controlling short-and long-term correlation effects. We propose a fast parameter estimation algorithm for afSm based on the Haar transform, and we demonstrate the performance of this parameter estimation algorithm with numerical simulations.

No abstract