Skip Decision and Reference Frame Selection for Low-Complexity H.264/AVC Surveillance Video Coding

H. 264/advanced video coding surveillance video encoders use the Skip mode specified by the standard to reduce bandwidth. They also use multiple frames as reference for motion-compensated prediction. In this paper, we propose two techniques to reduce the bandwidth and computational cost of static camera surveillance video encoders without affecting detection and recognition performance. A spatial sampler is proposed to sample pixels that are segmented using a Gaussian mixture model. Modified weight updates are derived for the parameters of the mixture model to reduce floating point computations. A storage pattern of the parameters in memory is also modified to improve cache performance. Skip selection is performed using the segmentation results of the sampled pixels. The second contribution is a low computational cost algorithm to choose the reference frames. The proposed reference frame selection algorithm reduces the cost of coding uncovered background regions. We also study the number of reference frames required to achieve good coding efficiency. Distortion over foreground pixels is measured to quantify the performance of the proposed techniques. Experimental results show bit rate savings of up to 94.5% over methods proposed in literature on video surveillance data sets. The proposed techniques also provide up to 74.5% reduction in compression complexity without increasing the distortion over the foreground regions in the video sequence.

Reduced complexity two stage vector quantization

We address the issue of complexity for vector quantization (VQ) of wide-band speech LSF (line spectrum frequency) parameters. The recently proposed switched split VQ (SSVQ) method provides better rate-distortion (R/D) performance than the traditional split VQ (SVQ) method, even at the requirement of lower computational complexity. but at the expense of much higher memory. We develop the two stage SVQ (TsSVQ) method, by which we gain both the memory and computational advantages and still retain good R/D performance. The proposed TsSVQ method uses a full dimensional quantizer in its first stage for exploiting all the higher dimensional coding advantages and then, uses an SVQ method for quantizing the residual vector in the second stage so as to reduce the complexity. We also develop a transform domain residual coding method in this two stage architecture such that it further reduces the computational complexity. To design an effective residual codebook in the second stage, variance normalization of Voronoi regions is carried out which leads to the design of two new methods, referred to as normalized two stage SVQ (NTsSVQ) and normalized two stage transform domain SVQ (NTsTrSVQ). These two new methods have complimentary strengths and hence, they are combined in a switched VQ mode which leads to the further improvement in R/D performance, but retaining the low complexity requirement. We evaluate the performances of new methods for wide-band speech LSF parameter quantization and show their advantages over established SVQ and SSVQ methods.

Fractional power dependence of rate on activation energy for reactions with very low internal barriers

The dynamics of reactions with low internal barriers are studied both analytically and numerically for two different models. Exact expressions for the average rate,kI, are obtained by solving the associated first passage time problems. Both the average rate constant, kI, and the numerically calculated long-time rate constant, kL, show a fractional power law dependence on the barrier height for very low barriers. The crossover of the reaction dynamics from low to high barrier is investigated.

Analysis-By-Synthesis Based Switched Transform Domain Split Vq Using Gaussian Mixture Model

Using analysis-by-synthesis (AbS) approach, we develop a soft decision based switched vector quantization (VQ) method for high quality and low complexity coding of wideband speech line spectral frequency (LSF) parameters. For each switching region, a low complexity transform domain split VQ (TrSVQ) is designed. The overall rate-distortion (R/D) performance optimality of new switched quantizer is addressed in the Gaussian mixture model (GMM) based parametric framework. In the AbS approach, the reduction of quantization complexity is achieved through the use of nearest neighbor (NN) TrSVQs and splitting the transform domain vector into higher number of subvectors. Compared to the current LSF quantization methods, the new method is shown to provide competitive or better trade-off between R/D performance and complexity.

A rate-one full-diversity low-complexity space-time-frequency block code (STFBC) for 4-Tx MIMO-OFDM

It is known that by employing space-time-frequency codes (STFCs) to frequency selective MIMO-OFDM systems, all the three diversity viz spatial, temporal and multipath can be exploited. There exists space-time-frequency block codes (STFBCs) designed using orthogonal designs with constellation precoder to get full diversity (Z.Liu, Y.Xin and G.Giannakis IEEE Trans. Signal Processing, Oct. 2002). Since orthogonal designs of rate one exists only for two transmit antennas, for more than two transmit antennas STFBCs of rate-one and full-diversity cannot be constructed using orthogonal designs. This paper presents a STFBC scheme of rate one for four transmit antennas designed using quasi-orthogonal designs along with co-ordinate interleaved orthogonal designs (Zafar Ali Khan and B. Sundar Rajan Proc: ISIT 2002). Conditions on the signal sets that give full-diversity are identified. Simulation results are presented to show the superiority of our codes over the existing ones.

An evaluation of rate-controlling obstacles for low-temperature deformation of zirconium

The low-temperature plastic flow of alpha-zirconium was studied by employing constantrate tensile tests and differential-stress creep experiments. The activation parameters, enthalpy and area, have been obtained as a function of stress for pure, as well as commercial zirconium. The activation area is independent of grain size and purity and falls to about 9b2 at high stresses. The deformation mechanism below about 700° K is found to be controlled by a single thermally activated process, and not a two-stage activation mechanism. Several dislocation mechanisms are examined and it is concluded that overcoming the Peierls energy humps by the formation of kink pairs in a length of dislocation is the rate-controlling mechanism. The total energy needed to nucleate a double kink is about 0.8 eV in pure zirconium and 1 eV in commercial zirconium

Dynamic programming based optimum non-uniform samples for speech reconstruction and coding

Non-uniform sampling of a signal is formulated as an optimization problem which minimizes the reconstruction signal error. Dynamic programming (DP) has been used to solve this problem efficiently for a finite duration signal. Further, the optimum samples are quantized to realize a speech coder. The quantizer and the DP based optimum search for non-uniform samples (DP-NUS) can be combined in a closed-loop manner, which provides distinct advantage over the open-loop formulation. The DP-NUS formulation provides a useful control over the trade-off between bitrate and performance (reconstruction error). It is shown that 5-10 dB SNR improvement is possible using DP-NUS compared to extrema sampling approach. In addition, the close-loop DP-NUS gives a 4-5 dB improvement in reconstruction error.

A Novel Construction of Complex Orthogonal Designs with Maximal Rate and Low-PAPR

Space-time block codes based on orthogonal designs are used for wireless communications with multiple transmit antennas which can achieve full transmit diversity and have low decoding complexity. However, the rate of the square real/complex orthogonal designs tends to zero with increase in number of antennas, while it is possible to have a rate-1 real orthogonal design (ROD) for any number of antennas.In case of complex orthogonal designs (CODs), rate-1 codes exist only for 1 and 2 antennas. In general, For a transmit antennas, the maximal rate of a COD is 1/2 + l/n or 1/2 + 1/n+1 for n even or odd respectively. In this paper, we present a simple construction for maximal-rate CODs for any number of antennas from square CODs which resembles the construction of rate-1 RODs from square RODs. These designs are shown to be amenable for construction of a class of generalized CODs (called Coordinate-Interleaved Scaled CODs) with low peak-to-average power ratio (PAPR) having the same parameters as the maximal-rate codes. Simulation results indicate that these codes perform better than the existing maximal rate codes under peak power constraint while performing the same under average power constraint.

On Computation of Minimum Distance of Linear Block Codes Above 1/2 Rate Coding

The minimum distance of linear block codes is one of the important parameter that indicates the error performance of the code. When the code rate is less than 1/2, efficient algorithms are available for finding minimum distance using the concept of information sets. When the code rate is greater than 1/2, only one information set is available and efficiency suffers. In this paper, we investigate and propose a novel algorithm to find the minimum distance of linear block codes with the code rate greater than 1/2. We propose to reverse the roles of information set and parity set to get virtually another information set to improve the efficiency. This method is 67.7 times faster than the minimum distance algorithm implemented in MAGMA Computational Algebra System for a (80, 45) linear block code.

Effect of strain rate on the high-temperature low-cycle fatigue properties of a nimonic PE-16 superalloy

Strain-rate effects on the low-cycle fatigue (LCF) behavior of a NIMONIC PE-16 superalloy have been evaluated in the temperature range of 523 to 923 K. Total-strain-controlled fatigue tests were per-formed at a strain amplitude of +/-0.6 pct on samples possessing two different prior microstructures: microstructure A, in the solution-annealed condition (free of gamma' and carbides); and microstructure B, in a double-aged condition with gamma' of 18-nm diameter and M23C6 carbides. The cyclic stress response behavior of the alloy was found to depend on the prior microstructure, testing temperature, and strain rate. A softening regime was found to be associated with shearing of ordered gamma' that were either formed during testing or present in the prior microstructure. Various manifestations of dynamic strain aging (DSA) included negative strain rate-stress response, serrations on the stress-strain hysteresis loops, and increased work-hardening rate. The calculated activation energy matched well with that for self-diffusion of Al and Ti in the matrix. Fatigue life increased with an increase in strain rate from 3 x 10(-5) to 3 x 10(-3) s-1, but decreased with further increases in strain rate. At 723 and 823 K and low strain rates, DSA influenced the deformation and fracture behavior of the alloy. Dynamic strain aging increased the strain localization in planar slip bands, and impingement of these bands caused internal grain-boundary cracks and reduced fatigue life. However, at 923 K and low strain rates, fatigue crack initiation and propagation were accelerated by high-temperature oxidation, and the reduced fatigue life was attributed to oxidation-fatigue interaction. Fatigue life was maximum at the intermediate strain rates, where strain localization was lower. Strain localization as a function of strain rate and temperature was quantified by optical and scanning electron microscopy and correlated with fatigue life.

High-Rate Space-Time Coded Large MIMO Systems: Low-Complexity Detection and Performance

Large MIMO systems with tens of antennas in each communication terminal using full-rate non-orthogonal space-time block codes (STBC) from Cyclic Division Algebras (CDA) can achieve the benefits of both transmit diversity as well as high spectral efficiencies. Maximum-likelihood (ML) or near-ML decoding of these large-sized STBCs at low complexities, however, has been a challenge. In this paper, we establish that near-ML decoding of these large STBCs is possible at practically affordable low complexities. We show that the likelihood ascent search (LAS) detector, reported earlier by us for V-BLAST, is able to achieve near-ML uncoded BER performance in decoding a 32x32 STBC from CDA, which employs 32 transmit antennas and sends 32(2) = 1024 complex data symbols in 32 time slots in one STBC matrix (i.e., 32 data symbols sent per channel use). In terms of coded BER, with a 16x16 STBC, rate-3/4 turbo code and 4-QAM (i.e., 24 bps/Hz), the LAS detector performs close to within just about 4 dB from the theoretical MIMO capacity. Our results further show that, with LAS detection, information lossless (ILL) STBCs perform almost as good as full-diversity ILL (FD-ILL) STBCs. Such low-complexity detectors can potentially enable implementation of high spectral efficiency large MIMO systems that could be considered in wireless standards.

X-Codes: A Low Complexity Full-Rate High-Diversity Achieving Precoder for TDD MIMO Systems

We consider a time division duplex multiple-input multiple-output (nt × nr MIMO). Using channel state information (CSI) at the transmitter, singular value decomposition (SVD) of the channel matrix is performed. This transforms the MIMO channel into parallel subchannels, but has a low overall diversity order. Hence, we propose X-Codes which achieve a higher diversity order by pairing the subchannels, prior to SVD preceding. In particular, each pair of information symbols is encoded by a fixed 2 × 2 real rotation matrix. X-Codes can be decoded using nr very low complexity two-dimensional real sphere decoders. Error probability analysis for X-Codes enables us to choose the optimal pairing and the optimal rotation angle for each pair. Finally, we show that our new scheme outperforms other low complexity precoding schemes.

High-rate analysis of channel-optimized vector quantization

High-rate analysis of channel-optimized vector quantizationThis paper considers the high-rate performance of channel optimized source coding for noisy discrete symmetric channels with random index assignment. Specifically, with mean squared error (MSE) as the performance metric, an upper bound on the asymptotic (i.e., high-rate) distortion is derived by assuming a general structure on the codebook. This structure enables extension of the analysis of the channel optimized source quantizer to one with a singular point density: for channels with small errors, the point density that minimizes the upper bound is continuous, while as the error rate increases, the point density becomes singular. The extent of the singularity is also characterized. The accuracy of the expressions obtained are verified through Monte Carlo simulations.

A Low ML-Decoding Complexity, Full-Diversity, Full-Rate MIMO Precoder

Precoding for multiple-input multiple-output (MIMO) antenna systems is considered with perfect channel knowledge available at both the transmitter and the receiver. For two transmit antennas and QAM constellations, a real-valued precoder which is approximately optimal (with respect to the minimum Euclidean distance between points in the received signal space) among real-valued precoders based on the singular value decomposition (SVD) of the channel is proposed. The proposed precoder is obtainable easily for arbitrary QAM constellations, unlike the known complex-valued optimal precoder by Collin et al. for two transmit antennas which is in existence for 4-QAM alone and is extremely hard to obtain for larger QAM constellations. The proposed precoding scheme is extended to higher number of transmit antennas on the lines of the E - d(min) precoder for 4-QAM by Vrigneau et al. which is an extension of the complex-valued optimal precoder for 4-QAM. The proposed precoder's ML-decoding complexity as a function of the constellation size M is only O(root M)while that of the E - d(min) precoder is O(M root M)(M = 4). Compared to the recently proposed X- and Y-precoders, the error performance of the proposed precoder is significantly better while being only marginally worse than that of the E - d(min) precoder for 4-QAM. It is argued that the proposed precoder provides full-diversity for QAM constellations and this is supported by simulation plots of the word error probability for 2 x 2, 4 x 4 and 8 x 8 systems.

Low-Delay, High-Rate Nonsquare Complex Orthogonal Designs

The maximal rate of a nonsquare complex orthogonal design for transmit antennas is 1/2 + 1/n if is even and 1/2 + 1/n+1 if is odd and the codes have been constructed for all by Liang (2003) and Lu et al. (2005) to achieve this rate. A lower bound on the decoding delay of maximal-rate complex orthogonal designs has been obtained by Adams et al. (2007) and it is observed that Liang's construction achieves the bound on delay for equal to 1 and 3 modulo 4 while Lu et al.'s construction achieves the bound for n = 0, 1, 3 mod 4. For n = 2 mod 4, Adams et al. (2010) have shown that the minimal decoding delay is twice the lower bound, in which case, both Liang's and Lu et al.'s construction achieve the minimum decoding delay. For large value of, it is observed that the rate is close to half and the decoding delay is very large. A class of rate-1/2 codes with low decoding delay for all has been constructed by Tarokh et al. (1999). In this paper, another class of rate-1/2 codes is constructed for all in which case the decoding delay is half the decoding delay of the rate-1/2 codes given by Tarokh et al. This is achieved by giving first a general construction of square real orthogonal designs which includes as special cases the well-known constructions of Adams, Lax, and Phillips and the construction of Geramita and Pullman, and then making use of it to obtain the desired rate-1/2 codes. For the case of nine transmit antennas, the proposed rate-1/2 code is shown to be of minimal delay. The proposed construction results in designs with zero entries which may have high peak-to-average power ratio and it is shown that by appropriate postmultiplication, a design with no zero entry can be obtained with no change in the code parameters.