Minimax codes for finite alphabets (Corresp.)

Minimax codes are designed for applications where the symbol probabilities are not known except for their order. A minimax code is defined by real-valued parameters/, (resembling the lengths of codewords) which minimize the maximum ratio(Sigma_{i}p_{i}l_{i})/(--Sigma_{i}p_{i}logp_{i})subject toSigma_{i}2^{-l_{i}}leq1over the set of monotonically nonincreasing distributions such thatp_{1}leq1/mforman integer greater than one.     

基于ROI分割和相干映射的裸手字母手势识别

提出了实时条件下裸手单目的手部定位和跟踪算法。结合改进的差分法提出了实用的手部约束条件，从每帧图像中提取手部图像，确定感兴趣区域（ROI，region of interesting），得到较为满意的手部分割结果，再对手部图像特征进行深入分析和提取，利用改进的相干映射算法（VCM，vector coherence mapping）进行跟踪，针对手的运动增加了约束，保证了顽健性。在这个基础上，提出了时间相关的运动预测模型，满足了实时性的要求，保证前后分析结果的一致性。实验结果证明，在不同光照和复杂背景下系统有最高达99％的识别率，与已有的系统相比，性能显著提高。     

Optimal encoding of discrete-time continuous-amplitude memoryless sources with finite output alphabets

A rate-distortion theory is introduced for the optimal encoding of stationary memoryless continuous-amplitude sources with a single-letter distortion measure and reproduction alphabets of a given finite size. The theory arises from a judicious approximation of the original continuous-input discrete-output problem by one with discrete input and output. A size-constrained output alphabet rate-distortion function is defined, its coding significance is established by coding theorems, and a convergent algorithm is presented for its evaluation. The theory is applied to Gaussian sources with squared-error distortion measure. Using the algorithm for the calculation of the new rate-distortion function in this case establishes the existence of codes which attain almost any desired rate between the rate-distortion bound and the optimum entropy-coded quantizer. Furthermore, one can closely approach the rate-distortion limit with a surprisingly small number of output levels. The calculation furnishes optimal output levels, output level probabilities, and other parameters necessary for a trellis coding simulation. The search algorithm represents the first use for asymmetric sources and distortion measures of a variation of a single stack algorithm proposed by Gallager. Carrying out the simulation at a rate of 1 bit per source symbol, codes are found with 4 and 64 output levels which attain distortions smaller than that of an optimum quantizer and close to the rate-distortion bound. Furthermore, these codes attain comparable or better performance with far less search effort than previous attempts with a continuous output alphabet.     

基于两空间核鉴别分析的人脸识别

指出了几种线性鉴别分析方法在处理小样本人脸识别问题时存在的不足,结合核方法的思想,提出了一种基于两空间核鉴别分析的人脸识别方法.首先使用KPCA方法在核变换后的特征空间中对样本进行处理;进面将变换后的类内散布矩阵分成非零空间和零空间进行鉴别向量确定和鉴别特征提取,最后将得到的两种鉴别特征融合,从而使用最近邻法进行分类....     

FIR system identification based on subspaces of a higher ordercumulant matrix

A new linear algebraic approach for identification of a nonminimum phase FIR system of known order using only higher order (>2) cumulants of the output process is proposed. It is first shown that a matrix formed from a set of cumulants of arbitrary order can be expressed as a product of structured matrices. The subspaces of this matrix are then used to obtain the parameters of the FIR system using a set of linear equations. Theoretical analysis and numerical simulation studies are presented to characterize the performance of the proposed methods     

LDPC codes over mixed alphabets

The notion of low-density parity-check (LDPC) codes over mixed alphabets is introduced. It is shown that these codes can provide good performance/complexity trade-offs compared to their single-alphabet counterparts     

Bandpass line codes for q-ary alphabets

The author considers codes with higher order spectral zeros at DC for q-ary alphabets. A recurrence relation is derived for the number of codewords that exist for a given length and moment, which enables the correction of some recently published tables for q=3 and q=5. Some minimum distance properties are determined for first order codes and any alphabet, and again some recently published results are corrected.<>     

基于Spark的分层子空间权重树随机森林算法

高维数据的很多特征与类别的相关性弱,影响了随机森林的分类正确率。针对原始随机森林算法在高维数据上的分类问题,提出了一种分层子空间权重树随机森林算法。同时,传统的单机模式无法满足高维数据计算效率的需求,因此利用开源集群计算框架Spark在内存缓存和迭代计算上的优势,将所提算法在Spark上实现。所提算法采用以决策树为单位的分层抽样来生成特征子空间,在提高单棵决策树性能的同时,保证决策树之间的多样性;并且采用权重树的集成策略,使分类能力强的树在集成过程中影响力更大。通过在Mnist和Gisette数据集上的实验结果表明,相比原始随机森林算法、TWRF算法以及分层子空间随机森林算法,所提算法具有更好的正确率,提高了泛化误差性能,可扩展性良好,能够有效分类高维数据。      

双辨别子空间高分辨距离像雷达目标识别

针对飞机目标的分类问题,提出了一种双辨别子空间高分辨距离像雷达目标识别方法.该方法首先依据Fisher准则导出距离像总散布矩阵的零空间中不含有辨别信息的结论,利用这一结论,对类间和类内散布矩阵进行预降维,降低了后续计算的复杂度.从全局的角度出发,基于类内散布矩阵零空间与非零空间所包含的辨别信息分别建立辨别子空间,实现对目标的特征提取.对ISAR实测飞机数据进行了分类,并与经典子空间方法进行比较,结果表明所提算法有效改善了目标识别性能.     

Spectral nulls and coding with large alphabets

Generalizations of the simple alternate mark inversion (AMI) line code that provide enhanced immunity to additive noise as well as spectral shaping are considered. The first technique is the extension of balanced disparity methods to multilevel signaling alphabets. This is a small step beyond simple codes such as AMI that are used to transmit binary PCM over twisted-pair cables. An important feature of this method (and later methods) is the use of suboptimal decoders. The most sophisticated technique used was Tomlinson (1971) filtering, and here it was possible to require a spectral null in the line code spectrum with a certain minimum width, and to minimize line code power subject to this requirement. This technique is compared to methods introduced by Marcus and Siegel (1987) in magnetic recording that provide spectral nulls at rational multiples of the symbol frequency. Theoretical possibilities, rather than techniques proven superior for a particular application, are addressed     

On identification capacity of infinite alphabets or continuous-timechannels

Two kinds of channels are considered: (1) discrete-time channel with additive noise, and (2) Poisson and white Gaussian (i.e., continuous-time) channels. For the type (1) channel there are given some sufficient conditions when the Shannon and identification capacities coincide. It is shown that the identification capacity of Poisson and Gaussian channels without bandwidth constraint is infinite. Conversely, for a white Gaussian channel with bandwidth constraint, the identification capacity coincides with the Shannon capacity     

Lowest density MDS codes over extension alphabets

Let F be a finite field and b be a positive integer. A construction is presented of codes over the alphabet F/sup b/ with the following three properties: i) the codes are maximum-distance separable (MDS) over F/sup b/, ii) they are linear over F, and iii) they have systematic generator and parity-check matrices over F with the smallest possible number of nonzero entries. Furthermore, for the case F=GF(2), the construction is the longest possible among all codes that satisfy properties i)-iii).     

Arbitrarily varying channels with general alphabets and states

Previous results of Csiszar and Narayan (1988, 1989, 1991) on the capacity of discrete arbitrarily varying channels with input and state constraints are extended to the case of arbitrary alphabets and state sets. For channels with scalar or vector inputs and additive interference consisting of a deterministic part and noise, both arbitrarily varying subject to power constraints, the capacity for deterministic codes is shown to be equal to the random coding capacity if the input power exceeds the power of the deterministic interference, and zero otherwise. Explicit capacity formulas are also given     

Universal compression of memoryless sources over unknown alphabets

It has long been known that the compression redundancy of independent and identically distributed (i.i.d.) strings increases to infinity as the alphabet size grows. It is also apparent that any string can be described by separately conveying its symbols, and its pattern-the order in which the symbols appear. Concentrating on the latter, we show that the patterns of i.i.d. strings over all, including infinite and even unknown, alphabets, can be compressed with diminishing redundancy, both in block and sequentially, and that the compression can be performed in linear time. To establish these results, we show that the number of patterns is the Bell number, that the number of patterns with a given number of symbols is the Stirling number of the second kind, and that the redundancy of patterns can be bounded using results of Hardy and Ramanujan on the number of integer partitions. The results also imply an asymptotically optimal solution for the Good-Turing probability-estimation problem.     

基于多维有限自动机的DFA改进算法

多个正则表达式规则编译成一个DFA(deter minister finite automata)时,会产生状态爆炸、存储急剧增加的现象。针对最严重的状态爆炸问题,从信息论的角度给出了解释,并提出多维数学模型,将冗余状态分为0维状态和1维状态,通过前者按照维度压缩,后者动态构建的方法将空间复杂度降到理论下界,并在此基础上提出多维有限自动机(MFA, multi-dimensional finite automata)。实验表明,MFA构造时间比XFA略少,比DFA、STT冗余压缩算法和Hybrid-FA降低了2~3个数量级;存储空间比XFA略高,比DFA、STT冗余压缩算法、mDFA、Hybrid-FA降低了1~2个数量级;匹配时间比DFA、Hybrid-FA略多,但是比XFA略少,比STT冗余压缩算法和mDFA降低了1~2个数量级。     

Variable-rate coding for classes of sources with generalized alphabets

Discrete-time source coding theorems are established for more general reproduction alphabets than allowed in previous results on coding for a class of sources subject to a fidelity constraint. The two different alphabets considered are metric spaces for which every closed bounded subset is compact and separable Hilbert spaces. Potential applications of the Hilbert space results to continuous-time source coding are discussed.     

On updating signal subspaces

The authors develop an algorithm for adaptively estimating the noise subspace of a data matrix, as is required in signal processing applications employing the `signal subspace' approach. The noise subspace is estimated using a rank-revealing QR factorization instead of the more expensive singular value or eigenvalue decompositions. Using incremental condition estimation to monitor the smallest singular values of triangular matrices, the authors can update the rank-revealing triangular factorization inexpensively when new rows are added and old rows are deleted. Experiments demonstrate that the new approach usually requires O(n²) work to update an n×n matrix, and that it accurately tracks the noise subspace     

基于有限差分法的方同轴线的研究

有限差分法是一种微分方法,是历史最悠久和理论最完整的数值分析方法。通过理论推导所求区域满足的偏微分方程和边界条件,然后在Yee网格中利用二阶有限差分法导出偏微分方程的差分格式,最后通过Matlab仿真出所求区域的电场分布和计算出方同轴线的特性阻抗。     

基于有限元的材料红外无损检测研究

通过有限元求解材料的二维导热模型,利用ANSYS仿真获取材料表面的温度分布,把表面温度计算值和仿真结果的误差平方和作为目标函数,借助于共轭梯度法来优化该目标函数.仿真结果验证了算法的可行性,可以达到红外无损检测的目的.研究表明:材料内部缺陷深度和材料的热导率是影响表面温差的重要因素,缺陷距离表面的深度越小,材料表面的温差越大;材料表面温差随热导率的增加呈现为先增大后减小的趋势.该方法可获得材料内部缺陷较为准确的位置和缺陷大小信息,为红外成像无损检测提供了可行的手段.     

Image encryption based on the finite field cosine transform

In this paper, a novel image encryption scheme is proposed. The technique involves two steps, where the finite field cosine transform is recursively applied to blocks of a given image. In the first step, the image blocks to be transformed result from the regular partition of subimages of the original image. The transformed subimages are regrouped and an intermediate image is constructed. In the second step, a secret-key determines the positions of the intermediate image blocks to be transformed. Besides complying with the main properties required by image encryption methods, the proposed scheme provides benefits related to computational complexity and encoding of the ciphered-images.