基于双点辅助的移动终端掌纹识别预处理算法

在移动终端掌纹采集和预处理过程中,针对难以 约束的手掌摆放位置和姿态、复杂场景、差异光照 和硬件资源有限等技术挑战,设计了一种基于指间根部位置的双点辅助拍摄方案,以辅 助点为中心的 正方形区域设定为限定框,并提出针对限定框的自适应高斯肤色模型的手掌分割方法。首先 选取限定框下 方矩形区域为训练子区域,建立高斯肤色模型,通过设定似然度阈值,分割限定框内手掌区 域;利用已分 割出的不完整手掌区域重新建立高斯肤色模型,再次设定肤色似然度阈值,准确分割出限定 框内的手部肤 色区域。实验对比后发现,与其他限定方法相比,本文方法降低了掌纹预处理算法的复杂度 ,改善了实时 性。与多种已有先进方案全面对比,本文方法不仅有效降低了复杂场景下手掌分割的失效率 ,在移动环境下还能够达到更高的认证精度。     

基于伽罗华域高斯列消元法的RS码盲识别

为解决高码率RS(Reed Solomon)码盲识别问题,提出了一种基于伽罗华域高斯列消元法的RS码盲识别方法。先利用矩阵秩的差值函数识别符号数及码长;再遍历此时符号数对应的本原多项式,对矩阵进行伽罗华域高斯列消元,并引入熵函数差值来识别本原多项式;最后求码字多项式的根,其中连续根即为生成多项式的根。该方法可以较好地识别RS码码长、生成多项式及本原多项式,并且避免了遍历符号数时多次进行伽罗华域傅里叶变换的繁琐过程。仿真结果表明,在误码率为3×10-3的情况下,对RS码的识别概率高于90%。     

基于高斯分布的自适应红外图像增强算法

针对红外图像信噪比和对比度低,结合红外图像直方图分布的特点,提出了一种基于高斯分布的自适应红外图像增强算法.该算法采用高斯分布函数平滑红外图像直方图中的主峰,以抑制红外图像背景和噪声的加强.算法中的高斯分布函数可根据原红外图像直方图的分布情况自适应产生,不需要人为设定,将所生成的高斯分布函数对原直方图进行转换后再进行直方图均衡处理,即可实现红外图像的增强.实验结果表明,该算法能有效地突显红外图像的目标,抑制其背景,具有较好的增强效果.     

基于广义高斯分布FDCT_Wrap和FSVM的虹膜识别

为了提高虹膜识别的准确率,提出了一种改进曲波变换的虹膜识别算法.首先对预处理后的虹膜图像进行Wrapping算法的快速离散曲波变换,提取不同尺度和不同方向的曲波子带系数矩阵的均值、方差和能量,然后利用广义高斯分布估算各子带的权值,为分类能力较强的特征向量赋予较大权值,构成虹膜图像的特征向量.最后采用模糊支持向量机和二叉决策树相结合的分类器进行匹配识别.采用UBIRIS和CASIA虹膜数据库对算法性能进行测试.实验结果表明,该算法能更好地提高虹膜识别准确率和效率,具有可行性.     

一种基于前端和后端模块精度优化的助听器系统芯片*

A hearing aid on-chip system based on accuracy optimized front- and back-end blocks is presented for enhancing the signal processing accuracy of the hearing aid. Compared with the conventional system, the accuracy optimized system is characterized by the dual feedback network and the gain compensation technique used in the front-andback-endblocks,respectively,soastoalleviatethenonlinearitydistortioncausedbytheoutputswing.By usingthetechnique,theaccuracyofthewholehearingaidsystemcanbesignificantlyimproved.Theprototypechip has been designed with a 0.13 m standard CMOS process and tested with 1 V supply voltage. The measurement results show that, for driving a 16 loudspeaker with a normalized output level of 300 mV p-p, the total harmonic distortion reached about60 dB, achieving at least three times reduction compared to the previously reported works. In addition, the typical input referred noise is only about 5 υV rms.     

基于高斯滤波器及费舍尔准则的特征提取方法

针对梅尔倒谱系数与翻转梅尔倒谱系数在语种识别应用中的不足,采用高斯滤波器代替三角滤波器,提出一种新的梅尔倒谱系数提取方法,解决传统梅尔倒谱系数提取中邻近滤波器相关性较弱的问题,并结合Fisher准则构造出最优混合特征参数,采用高斯混合模型分别对不同混合特征进行语种识别。实验结果表明:基于高斯滤波器及Fisher准则的改进梅尔倒谱系数混合特征参数作为语种识别特征具有较高的识别准确率。     

基于免疫克隆高斯过程隐变量模型的SAR目标特征提取与识别

作为一种非线性维数约减算法,高斯过程隐变量模型(Gaussian process latent variable model,GPLVM)由于其适合处理小样本、高维数据,因而在模式识别、计算机视觉等领域得到了广泛应用.基于此,提出一种基于改进GPLVM的SAR图像目标特征提取及自动识别方法,其中利用改进的GPLVM进行特征提取,高斯过程分类进行目标识别.传统GPLVM使用共轭梯度法对似然函数进行优化,为避免梯度估值易受噪声干扰、步长对算法影响严重等缺点,提出基于免疫克隆选择算法的GPLVM,利用其具有快速收敛到全局最优的特性提高算法性能.实验结果表明,该算法不仅降低了特征维数,且提高了识别精度,从而验证了算法用于SAR图像目标识别的有效性.     

DWDM系统中基于高斯噪声模型的调制格式识别研究

采用高斯噪声(GN)模型,研究了密集波分复用(DWD M)系统中不同调制格式的识别性能。调制格式共有23种, 除包含MASK、MPSK和MQAM等规则调制格式外,也包含二维最优星座图、高阶APSK星座图等经 过性 能优化的调制格式。针对MASK、MQAM和MAPSK调制识别的部分特征参数随信噪比(SNR)变化的问题,提出 了基于训练序列的阈值优化方法,使其低SNR下的识别率提高至93%以上。研究了DWDM系统的色散(CD)、 非线性及传输距离对识别率的影响,仿真分析了不同调制格式的识别性能。仿真结果表明, MAPSK是最 优的调制类型,识别率高于99%;二维最优星座图是次优的调制类型 ,其识别率高于93%。     

Block-based histogram of optical flow for isolated sign language recognition

In this paper, we propose a block-based histogram of optical flow (BHOF) to generate hand representation in sign language recognition. Optical flow of the sign language video is computed in a region centered around the location of the detected hand position. The hand patches of optical flow are segmented into M spatial blocks, where each block is a cuboid of a segment of a frame across the entire sign gesture video. The histogram of each block is then computed and normalized by its sum. The feature vector of all blocks are then concatenated as the BHOF sign gesture representation. The proposed method provides a compact scale-invariant representation of the sign language. Furthermore, block-based histogram encodes spatial information and provides local translation invariance in the extracted optical flow. Additionally, the proposed BHOF also introduces sign language length invariancy into its representation, and thereby, produce promising recognition rate in signer independent problems.     

广义高斯信源的自适应编码研究

首先介绍了当前最新的几种视频编码标准并且进行了比较,熵编码是每一个视频编码标准必须认真研究的课题,为了减少由于失配所带来的效率损失,本文提出了一种自适应编码技术:自适应指数哥伦布码,并与自适应算术编码进行了比较.分析和仿真都表明即使信源特性在大范围内发生变化,自适应指数哥伦布码对于保持高的编码效率是足够稳健的(90%以上的情况),同时保持了指数哥伦布码和哥伦布-莱期码的简洁性.     

Architectures for language recognition

Two distinct chip architectures for language recognition are presented. The first allows a recogniser for any regular expression to be constructed by simply assembling predefined cells into a linear array. The cells correspond to the symbols which can occur in a regular expression, and the appropriate cells are placed in a line in the same order as the expression. The recogniser is then correct by construction, and all inter-cell connections are made automatically. The second architecture can be programmed to recognise any LL(1) context-free language by loading it with the appropriate grammer rules. The grammar rules are embodied within an array of identical cells, which, in conjunction with a standard LIFO stack, form the recognition hardware.     

基于改进高斯模型的车流量检测算法

提出了一种基于改进单高斯模型的车辆检测与流量统计算法,该算法采用改进的单高斯模型对移动目标进行检测,然后选用HSV颜色空间抑制阴影,提高了目标提取的准确率,最后,按车道分别设置相应的虚拟区域,以实现车流量的统计工作.为验证算法的有效性,在标准测试视频上进行了对比实验,实验结果表明,该算法能够快速地提取车辆目标,且具有较高的车辆识别率,有一定的实用价值.     

Multi-view motion modelled deep attention networks (M2DA-Net) for video based sign language recognition

Currently, video-based Sign language recognition (SLR) has been extensively studied using deep learning models such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs). In addition, using multi view attention mechanism along with CNNs could be an appealing solution that can be considered in order to make the machine interpretation process immune to finger self-occlusions. The proposed multi stream CNN mixes spatial and motion modelled video sequences to create a low dimensional feature vector at multiple stages in the CNN pipeline. Hence, we solve the view invariance problem into a video classification problem using attention model CNNs. For superior network performance during training, the signs are learned through a motion attention network thus focusing on the parts that play a major role in generating a view based paired pooling using a trainable view pair pooling network (VPPN). The VPPN, pairs views to produce a maximally distributed discriminating features from all the views for an improved sign recognition. The results showed an increase in recognition accuracies on 2D video sign language datasets. Similar results were obtained on benchmark action datasets such as NTU RGB D, MuHAVi, WEIZMANN and NUMA as there is no multi view sign language dataset except ours.     

基于自适应Volterra的高斯噪声图像滤波算法

针对高斯噪声图像滤波问题,提出一种可以保留细节的自适应Volterra-Volterra滤波方法。为了解决二阶Volterra滤波器核系数确定难、数量多等问题,提出了一种自适应建核结构及算法。通过图像处理实验,与其它滤波算法进行了细节度保留、均方差(MSE)和信噪比(SNR)等指标的比较,结果表明,本文方法优于进行对比的其它滤波方法。     

自适应高斯混合模型及说话人识别应用

高斯混合模型采用固定混合数结构的建模方法并不符合说话人语音特征分布的多样性,从而出现过拟合或者欠拟合的情况并影响系统的识别性能。提出一种混合数可变的自适应高斯混合模型并将其应用于说话人识别。模型训练中根据说话人语音特征参数分布的聚类特性,采用吸收合并与分裂机制动态调整混合数以获得更加精确的拟合性能,提高系统识别率。实验结果显示,在特征参数MFCC和BFCC（Bilinear Frequency Cepstrum Coefficients）下相对误识率分别下降了41.41%和22.21%。     

Radar HRRP recognition based on the minimum Kullback-Leibler Distance criterion

To relax the target aspect sensitivity and use more statistical information of the High Range Resolution Profiles （HRRPs）, in this paper, the average range profile and the variance range profile are extracted together as the feature vectors for both training data and test data representation. And a decision rule is established for Automatic Target Recognition （ATR） based on the minimum Kullback-Leibler Distance （KLD） criterion. The recognition performance of the proposed method is comparable with that of Adaptive Gaussian Classifier （AGC） with multiple test HRRPs, but the proposed method is much more computational efficient. Experimental results based on the measured data show that the minimum KLD classifier is effective.     

基于弯曲可变电阻的手语识别系统设计

文章结合陀螺仪(GYRO)、加速度计(ACC)、弯曲可变电阻三者收集到的手势动作信息,提出一种融合多传感器信息的手势识别方法,以提高训练及识别的速度。首先执行运动窗口检测,然后平滑滤波,将手势加速信号预处理为帧,最后构建离散隐马尔可夫模型以实现手势识别。     

最大边缘准则图形嵌入在人脸识别中的应用

局部线性嵌入是一种有效地非线性维数约减方法,它能保持降维后的数据与原空间有相同的拓扑关系。但是这种方法在降维处理、可视化以及数据分类方面应用不是很广泛,针对上述问题,提出了一种新的、有效的降维以及数据分类方法——基于最大边缘准则图形嵌入方法。该方法首先构建最近邻关系图聚合数据点之间的最近邻样本,同时最大化类间间隔,保证不同类之间数据可分性大,从而更好地实现数据分类。最后,该方法的有效性分别在ORL及Yale两大人脸库上得到了验证。