An Image Segmentation Algorithm Based on a Local Region Conditional Random Field Model

To reduce the computation cost of a combined probabilistic graphical model and a deep neural network in semantic segmentation, the local region condition random field (LRCRF) model is investigated which selectively applies the condition random field (CRF) to the most active region in the image. The full convolutional network structure is optimized with the ResNet-18 structure and dilated convolution to expand the receptive field. The tracking networks are also improved based on SiameseFC by considering the frame relations in consecutive-frame traffic scene maps. Moreover, the segmentation results of the greyscale input data sets are more stable and effective than using the RGB images for deep neural network feature extraction. The experimental results show that the proposed method takes advantage of the image features directly and achieves good real-time performance and high segmentation accuracy.

The nonparametric estimation of long memory spatio-temporal random field models

This paper considers the local linear estimation of a multivariate regression function and its derivatives for a stationary long memory(long range dependent) nonparametric spatio-temporal regression model.Under some mild regularity assumptions, the pointwise strong convergence, the uniform weak consistency with convergence rates and the joint asymptotic distribution of the estimators are established. A simulation study is carried out to illustrate the performance of the proposed estimators.

Spin-1 Blume-Capel model with longitudinal random crystal and transverse magnetic fields:A mean-field approach

The spin-1 Blume–Capel model with transverse and longitudinal external magnetic fields h, in addition to a longitudinal random crystal field D, is studied in the mean-field approximation. It is assumed that the crystal field is either turned on with probability p or turned off with probability 1 p on the sites of a square lattice. Phase diagrams are then calculated on the reduced temperature crystal field planes for given values of γ=Ω/J and p at zero h. Thus, the effect of changing γ and p are illustrated on the phase diagrams in great detail and interesting results are observed.

Longitudinal-Random-Field Mixed Ising Model with Arbitrary Spins

混合 Ising 模型由任意的纺纱组成珍视的 longitudinal-random-field 被一个有效领域理论的使用与关联(水蜥) 学习了。有混合旋转的系统的阶段图：= 1/2， S = 1；= 1/2， S = 3/2 被阴谋。不仅在 T = 的断绝 0 K，当纵的领域是 trimodal 时，被发现散布了，而且 tricritical 行为在在 bimodal 和纵的领域的 trimodal 分布之间的这些阶段图被观察，它与单个纺纱的不同。tricritical 点的外观独立于协作数字和旋转价值。

Modified Maximum Likelihood Estimation of the Spatial Resolution for the Elliptical Gamma Camera SPECT Imaging Using Binary Inhomogeneous Markov Random Fields Models

In this work a complete approach for estimation of the spatial resolution for the gamma camera imaging based on the [1] is analyzed considering where the body distance is detected (close or far way). The organ of interest most of the times is not well defined, so in that case it is appropriate to use elliptical camera detection instead of circular. The image reconstruction is presented which allows spatially varying amounts of local smoothing. An inhomogeneous Markov random field (M.r.f.) model is described which allows spatially varying degrees of smoothing in the reconstructions and a re-parameterization is proposed which implicitly introduces a local correlation structure in the smoothing parameters using a modified maximum likelihood estimation (MLE) denoted as one step late (OSL) introduced by [2].

Magnetic-resonance image segmentation based on improved variable weight multi-resolution Markov random field in undecimated complex wavelet domain

To solve the problem that the magnetic resonance(MR)image has weak boundaries,large amount of information,and low signal-to-noise ratio,we propose an image segmentation method based on the multi-resolution Markov random field(MRMRF)model.The algorithm uses undecimated dual-tree complex wavelet transformation to transform the image into multiple scales.The transformed low-frequency scale histogram is used to improve the initial clustering center of the K-means algorithm,and then other cluster centers are selected according to the maximum distance rule to obtain the coarse-scale segmentation.The results are then segmented by the improved MRMRF model.In order to solve the problem of fuzzy edge segmentation caused by the gray level inhomogeneity of MR image segmentation under the MRMRF model,it is proposed to introduce variable weight parameters in the segmentation process of each scale.Furthermore,the final segmentation results are optimized.We name this algorithm the variable-weight multi-resolution Markov random field(VWMRMRF).The simulation and clinical MR image segmentation verification show that the VWMRMRF algorithm has high segmentation accuracy and robustness,and can accurately and stably achieve low signal-to-noise ratio,weak boundary MR image segmentation.

STUDIES ON THERMODYNAMICAL PROPERTIES OF RANDOM-BOND ISING MODEL IN A TRANSVERSE FIELD WITH CEA AND DPIR

A effective approximate scheme which is combined by cluster with the discrelized path-integral representation (DPIR) is used in the study on the random-bond Ising model in a transverse field (RTIM). The critical thermodynamical properties, such as the critical temperature, the critical transverse field, the average magnetization ,the susceptibility and the special heat atc.. are calculated, And some results have been improved.

Simulation of Random Crack Generation in Concrete Members with Uniform Stress Fields

The randomness of strength and deformation of concrete material is serious and should be considered both in theoretical analyses such as Finite Element Methods and engineering practice, specially for those structural members with a uniform stress field, where stresses or strains are approximately the same under loading. A mathematical ap- proach of producing a series of random variables of the ultimate tensile strain in concrete is proposed to describe the randomness ofconcrete deformation. With reinforced concrete finite elements a real model calculation method is found for the randomness of initial cracks determined by a minimum tension strain within the uniform stress fields of concrete members. The proposed methods in our paper have as aim to improve the existing method used by FEM and other rela- tive approaches, which normally pay less attention to randomness with consequences that may possibly differ from testing or practice. The method and sample computation as indicated is meaningful and comply with testing and engi- neering practice.

Reservoir lithology stochastic simulation based on Markov random fields

Markov random fields(MRF) have potential for predicting and simulating petroleum reservoir facies more accurately from sample data such as logging, core data and seismic data because they can incorporate interclass relationships. While, many relative studies were based on Markov chain, not MRF, and using Markov chain model for 3D reservoir stochastic simulation has always been the difficulty in reservoir stochastic simulation. MRF was proposed to simulate type variables(for example lithofacies) in this work. Firstly, a Gibbs distribution was proposed to characterize reservoir heterogeneity for building 3-D(three-dimensional) MRF. Secondly, maximum likelihood approaches of model parameters on well data and training image were considered. Compared with the simulation results of MC(Markov chain), the MRF can better reflect the spatial distribution characteristics of sand body.

Optimization by Estimation of Distribution with DEUM Framework Based on Markov Random Fields

This paper presents a Markov random field （MRP） approach to estimating and sampling the probability distribution in populations of solutions. The approach is used to define a class of algorithms under the general heading distribution estimation using Markov random fields （DEUM）. DEUM is a subclass of estimation of distribution algorithms （EDAs） where interaction between solution variables is represented as an undirected graph and the joint probability of a solution is factorized as a Gibbs distribution derived from the structure of the graph. The focus of this paper will be on describing the three main characteristics of DEUM framework, which distinguishes it from the traditional EDA. They are： 1） use of MRF models, 2） fitness modeling approach to estimating the parameter of the model and 3） Monte Carlo approach to sampling from the model.

Probabilistic stability analyses of undrained slopes by 3D random fields and finite element methods

A long slope consisting of spatially random soils is a common geographical feature. This paper examined the necessity of three-dimensional(3 D) analysis when dealing with slope with full randomness in soil properties. Although 3 D random finite element analysis can well reflect the spatial variability of soil properties, it is often time-consuming for probabilistic stability analysis. For this reason, we also examined the least advantageous(or most pessimistic) cross-section of the studied slope. The concept of"most pessimistic" refers to the minimal cross-sectional average of undrained shear strength. The selection of the most pessimistic section is achievable by simulating the undrained shear strength as a 3 D random field. Random finite element analysis results suggest that two-dimensional(2 D) plane strain analysis based the most pessimistic cross-section generally provides a more conservative result than the corresponding full 3 D analysis. The level of conservativeness is around 15% on average. This result may have engineering implications for slope design where computationally tractable 2 D analyses based on the procedure proposed in this study could ensure conservative results.

Phase diagrams of the spin-2 Ising model in the presence of a quenched diluted crystal field distribution

We have investigated the random crystal field effects on the phase diagrams of the spin-2 Blume-Capel model for a honeycomb lattice using the effective-field theory with correlations. To do so, the thermal variations of magnetization are studied via calculating the phase diagrams of the model. We have found that the model displays both second-order and first-order phase transitions in addition to the tricritical and isolated points. Reentrant behavior is also observed for some appropriate values of certain system parameters. Besides the usual ground-state phases of the spin-2 model including ±2, ~1, and 0, we have also observed the phases ±3/2 and ±1/2, which are unusual for the spin-2 case.

基于改进DeeplabV3+的水面多类型漂浮物分割方法研究

【目的】为解决传统图像处理方法鲁棒性差、常用深度学习检测方法无法准确识别大片漂浮物的边界等问题,【方法】提出一种基于改进DeeplabV3+的水面多类型漂浮物识别的语义分割方法,提高水面漂浮的识别能力。对所收集实际水面漂浮物进行分类,采用自制数据集进行对比试验。算法选择xception网络作为主干网络以获得初步漂浮物特征,在加强特征提取网络部分引入注意力机制以强调有效特征信息,在后处理阶段加入全连接条件随机场模型,将单个像素点的局部信息与全局语义信息融合。【结果】对比图像分割性能指标,改进后的算法mPA(Mean Pixel Accuracy)提升了5.73%,mIOU(Mean Intersection Over Union)提升了4.37%。【结论】相比于其他算法模型,改进后的DeeplabV3+算法对漂浮物特征的获取能力更强,同时能获得丰富的细节信息以更精准地识别多类型水面漂浮物的边界与较难分类的漂浮物,在对多个水库场景测试后满足实际水域环境中漂浮物检测的需求。

基于局部Transformer的泰语分词和词性标注联合模型

泰语分词和词性标注任务二者之间存在高关联性,已有研究表明将分词和词性标注任务进行联合学习可以有效提升模型性能,为此,提出了一种针对泰语拼写和构词特点的分词和词性标注联合模型。针对泰语中字符构成音节,音节组成词语的特点,采用局部Transformer网络从音节序列中学习分词特征;考虑到词根和词缀等音节与词性的关联,将用于分词的音节特征融入词语序列特征,缓解未知词的词性标注特征缺失问题。在此基础上,模型采用线性分类层预测分词标签,采用线性条件随机场建模词性序列的依赖关系。在泰语数据集LST20上的试验结果表明,模型分词F1、词性标注微平均F1和宏平均F1分别达到96.33%、97.06%和85.98%,相较基线模型分别提升了0.33%、0.44%和0.12%。

基于BERT-BiLSTM-CRF模型的畜禽疫病文本分词研究

针对畜禽疫病文本语料匮乏、文本内包含大量疫病名称及短语等未登录词问题,提出了一种结合词典匹配的BERT-BiLSTM-CRF畜禽疫病文本分词模型。以羊疫病为研究对象,构建了常见疫病文本数据集,将其与通用语料PKU结合,利用BERT(Bidirectional encoder representation from transformers)预训练语言模型进行文本向量化表示;通过双向长短时记忆网络(Bidirectional long short-term memory network,BiLSTM)获取上下文语义特征;由条件随机场(Conditional random field,CRF)输出全局最优标签序列。基于此,在CRF层后加入畜禽疫病领域词典进行分词匹配修正,减少在分词过程中出现的疫病名称及短语等造成的歧义切分,进一步提高了分词准确率。实验结果表明,结合词典匹配的BERT-BiLSTM-CRF模型在羊常见疫病文本数据集上的F1值为96.38%,与jieba分词器、BiLSTM-Softmax模型、BiLSTM-CRF模型、未结合词典匹配的本文模型相比,分别提升11.01、10.62、8.3、0.72个百分点,验证了方法的有效性。与单一语料相比,通用语料PKU和羊常见疫病文本数据集结合的混合语料,能够同时对畜禽疫病专业术语及疫病文本中常用词进行准确切分,在通用语料及疫病文本数据集上F1值都达到95%以上,具有较好的模型泛化能力。该方法可用于畜禽疫病文本分词。

基于本体驱动的航空情报表格信息结构化研究

航空资料汇编是国际民航组织推荐的呈现各国航空信息的主要载体,其中以表格数据形式汇总了大量航空数据与航空运行限制信息。为实现航空汇编资料的智能查询,以及对航空资料汇编中静态数据的挖掘与利用,需要对航空汇编资料中的表格信息予以特征提取与结构化处理。将航空资料汇编中表格信息作为研究对象,提出了一种基于本体驱动的航空情报表格信息结构化抽取方法。首先构建航空情报领域信息的本体框架,实现对领域知识统一规范的描述;其次,利用Document AI对表格文档的布局结构进行研究与预处理,并利用随机森林算法与条件随机场模型进行特征实体提取验证与分析。实验结果表明,所提方法能够有效提取航空情报表格中的特征实体,为航空情报领域静态数据深入挖掘提供参考。

结合马尔可夫随机场和混合模型的海岸线提取

为了高效且准确地实现基于遥感影像的海岸线提取,提出一种结合马尔可夫随机场和混合模型的合成孔径雷达(synthetic aperture radar,SAR)影像海岸线提取算法。该算法以统计模型理论为研究基础,考虑SAR影像中同一地物像素反射强度的统计分布具有非对称和重尾的统计特性,利用伽马混合模型建立SAR影像内像素强度的概率分布。为了建模像素的空间相关性,采用马尔可夫随机场构建伽马混合模型的组分权重概率分布以克服SAR影像相干斑噪声的影响。结合马尔可夫随机场和伽马混合模型构建出SAR影像海陆分割模型,通过最大期望方法估计模型参数以实现准确的海陆分割,进而实现海岸线提取。在Sentinel-1卫星SAR影像上进行海岸线提取实验,实验结果表明该算法可实现准确的海岸线提取。

双模随机晶体场对混合自旋1/2和自旋1纳米管上相变的影响

本文利用有效场理论研究了自旋1/2和自旋1混合Blume-Capel模型在具有双模随机晶体场的圆柱形Ising纳米管上的磁化和相变.通过数值计算,我们得到了随温度和随机晶体场参数变化的相图和磁化强度.结果表明:(1)改变晶体场的概率和比例,双模随机晶体场可以描述不同掺杂原子对自旋的作用;(2)对于一定的概率值、负或正的晶体场和晶体场的比例值都存在临界点;(3)系统显示多种相变温度,一阶相变和二阶相变.

基于Voronoi图与条件随机场的自然场景文本检测方法

在自然场景中准确有效地检测文本是一项艰巨的任务,故提出一种基于条件随机场(CRF)框架的场景文本检测方法。通过利用贝叶斯推断估计文本极大值区域的置信度作为一元成本项,通过使用维诺图(Voronoi图)来构建CRF空间邻域信息,从而构建图模型,通过最大流算法最小化成本函数区分文本与非文本标记;利用字符的几何特性通过聚类方法聚合成行。实验结果表明,该算法比传统基于最大稳定极值区域(MSER)算法性能有所提高,自然场景文本检测正确率能达到87%。

基于BERT-BiLSTM-CRF模型的油气领域命名实体识别

针对油气领域知识图谱构建过程中命名实体识别使用传统方法存在实体特征信息提取不准确、识别效率低的问题,提出了一种基于BERT-BiLSTM-CRF模型的命名实体识别研究方法。该方法首先利用BERT(bidirectional encoder representations from transformers)预训练模型得到输入序列语义的词向量;然后将训练后的词向量输入双向长短期记忆网络(bi-directional long short-term memory,BiLSTM)模型进一步获取上下文特征;最后根据条件随机场(conditional random fields,CRF)的标注规则和序列解码能力输出最大概率序列标注结果,构建油气领域命名实体识别模型框架。将BERT-BiLSTM-CRF模型与其他2种命名实体识别模型(BiLSTM-CRF、BiLSTM-Attention-CRF)在包括3万多条文本语料数据、4类实体的自建数据集上进行了对比实验。实验结果表明,BERT-BiLSTM-CRF模型的准确率(P)、召回率(R)和F_(1)值分别达到91.3%、94.5%和92.9%,实体识别效果优于其他2种模型。