Two algorithms for block motion estimation that produce performance similar to of exhaustive search but with computation reduced by a factor 8 or 16 are presented. The based on motion-field and pixel subsampling. A subsampled field is first determined estimating the vectors fraction blocks. these blocks using only pixels at any searched location alternating subsampling patterns locations. then interpolated so vector each pixels. more robust than previously proposed fast both can easily be...

A technique using boundary matching to compensate for lost or erroneously received motion vectors in motion-compensated video coding is proposed. This technique, called the algorithm, produces noticeably better results than those reported previously. It first assumed that displaced frame differences have no error. Then, this assumption relaxed by proposing an algorithm (the extended EBMA) which can recover both missing and vectors. The resulting images obtained these methods some other are...

A multiple resolution algorithm is presented for segmenting images into regions with differing statistical behavior. In addition, an developed determining the number of statistically distinct in image and estimating parameters those regions. Both algorithms use a causal Gaussian autoregressive model to describe mean, variance, spatial correlation textures. Together, can be used perform unsupervised texture segmentation. The segmentation first segments at coarse then progresses finer...

An analysis of kinoform image reconstruction error is presented. This considers the effects introduced by approximation and quantization plotting. The measure developed applied to a proposed method for computing kinoforms. Numerical results indicate that images produced this offer considerable reduction in when compared with from kinoforms made random phase method.

This paper calculates the error at output of a digital filter using floating-point arithmetic operations due to roundoff accumulation and input quantization. Expressions are derived for mean square each three canonical forms realization: direct, cascade, parallel. Upper bounds that useful special class filters given. The theoretical result is found be in good agreement with experimental result.

Summary Production optimization led by computing intelligence can greatly improve oilfield economic effectiveness. However, it is confronted with huge computational challenge because of the expensive black-box objective function and high-dimensional design variables. Many low-fidelity methods based on simplified physical models or data-driven have been proposed to reduce evaluation costs. These approximate global fitness landscape a certain extent, but difficult ensure accuracy correlation...

Abstract High temperature‐proton exchange membrane fuel cells (HT‐PEMFC) call for ionomers with low humidity dependence and elevated‐temperature resistance. Traditional perfluorosulfonic acid (PFSA) encounter challenges in meeting these stringent requirements. Herein, this study reports a perfluoroimide multi‐acid (PFMA) ionomer dual active centers achieved through the incorporation of sulfonimide phosphonic groups into side chain. The fluorocarbon skeleton chain structure facilitate...

Digital watermarking has been proposed as the means for copyright protection of multimedia data. Many existing schemes focused on robust to mark an image invisibly without really addressing ends these schemes. This paper first discusses some scenarios in which many current fail resolve rightful ownership image. The key problems are then identified, and crucial requirements a valid invisible watermark detection discussed. In particular, we show that, particular application resolving using...

Stochastic gradient adaptive filtering algorithms using variable step sizes are investigated. The variable-step-size algorithm improves the convergence rate while sacrificing little in steady-state error. Expressions describing of mean and mean-squared values coefficients developed used to calculate mean-square-error evolution. initial error also performance is studied when a power-of-two quantizer used, finite-word-length effects considered. analytical results verified with simulations...

The authors show that by cascading two direct-form FIR filters, each with coefficients are the sum or difference of power-of-two terms, it is possible to achieve very small peak ripple. An iterative equalization strategy used in design cascaded filter. success method depends on initial prototype filter being used. Two equally effective methods presented for selecting filter; yields a final good roundoff noise property.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML"...

The present paper considers discrete time analyses of firstand second-order digital phase lock loops. These loops are characterized by the fact that they track zero crossings incoming signal; consequently, sampling intervals nonuniform. firstorder loop is analyzed for step and frequency inputs; mean to skip cycle also considered. For input, approximate expressions obtained steady-state error probability density variance, second which leads directly a theoretical prediction threshold. input....

The previous treatments for strange quark matter in the mass-density-dependent model have unreasonable vacuum limits. We provide a method to obtain mass parametrizations and give self-consistent thermodynamic treatment which includes MIT bag as an extreme. In this treatment, bulk still has possibility of absolute stability. However, behavior sound velocity is opposite findings.

Reduced images formed from the collection of DC coefficients in intra-coded DCT compressed video retain a "global" feature which is useful for processing purposes. This paper examines direct reconstruction such motion-compensated P-frames and B-frames MPEG video. Exact expressions are derived approximations given computational speedup. The analysis experimental results show that yield very close to true ones.

A common framework for rapid scene analysis detecting changes in compressed motion JPEG and MPEG videos is proposed. We develop algorithms to detect both abrupt gradual changes. The operate directly on the DC sequence which can be easily extracted from video without decompression. images capture most of essential "global" information, but a small fraction original data size. Operating these offers significant computation savings. Experimental results show that proposed are fast effective transitions.

Previous work has shown that approximation of digital filter coefficients using sums signed power-of-two terms yields significant area/speed advantages in custom implementations, at the expense a slight frequency response deterioration. The completeness, uniqueness, and resolving power powers-of-two representations are studied, circuits for extracting prescribed number whose sum is closest to given integer presented. Examples implementation these CMOS process given.< <ETX...

When timing jitter is present in a sampled-data system, it important to be able calculate the error introduced by jittered sampler. However, detailed statistical description of not usually available allow determination and optimum filters. This paper calculates bounds for such error. For single channel sampling, are obtained cases where independent from sample correlated. An bound special case band-pass sampling using two samplers also derived. In all cases, signal assumed properly...