The Chinese character library is one of the important data structures in the Chinese information Processing system.The behavior of the whole system depends directly on the reasonableness of design for its structure.Th...The Chinese character library is one of the important data structures in the Chinese information Processing system.The behavior of the whole system depends directly on the reasonableness of design for its structure.This paper expounds the structures of RAM-based Chinese character libraries,static and dynamic ,The paper offers a descriptive method for this behavior and inquires into some algorithms related to the structures mentioned above.展开更多
The segmentation of individual words into characters is a vital process in handwritten character recognition systems. In this paper, a novel approach is proposed to segment handwritten Arabic text (words). We consider...The segmentation of individual words into characters is a vital process in handwritten character recognition systems. In this paper, a novel approach is proposed to segment handwritten Arabic text (words). We consider the “Naskh” font style. The segmentation algorithm employs seven agents in order to detect regions where segmentation is illegal. Feature points (end points) are extracted from the remaining regions of the word-image. Initially, the middle of every two successive end points is considered as a candidate segmentation point based on a set of rules. The experimental results are very promising as we achieved a success rate of 86%.展开更多
A new millet (Setaria italica Beauv) variety, super early-mature millet No.1, was bred by means of gene bank breedingmethod of target characters. This variety has the following outstanding characters. (1) Super early-...A new millet (Setaria italica Beauv) variety, super early-mature millet No.1, was bred by means of gene bank breedingmethod of target characters. This variety has the following outstanding characters. (1) Super early-mature. This varietyonly needs 1550C effective accumulated temperature and can normally maturate in the Bashang Region in Hebei Provinceof Chi na, which can break through the limit zone of millet cultivation and move the cultivation zone northward greatly. (2)Multi-spikes, in addition to the effect tilling at the top, the nodes in the low-middle part also can produce spikes. (3) Sweetstem have high sugar content. The contents of whole-sugar, soluable sugar and deoxidized sugar are 74.8, 200.5, 237.2%higher than the regular varieties respectively. (4) High gross protein content. The content of gross protein is higher thanthe regular varieties by 3.9-30.4%. (5)Changeable grain color. The grain color of super early-mature millet No.1 is red inShijiazhuang, but yellow in the Bashang region. In addition, this variety is characterized by good quality, high yield, andgood synthetic traits展开更多
脱机手写中文字符识别(handwritten Chinese character recognition,HCCR)在计算机视觉领域一直是一个巨大的挑战。相比传统方法,基于深度学习的网络通过训练大量数据在识别任务中取得了差异化的效果,但识别效果依旧处于发展过程中。基...脱机手写中文字符识别(handwritten Chinese character recognition,HCCR)在计算机视觉领域一直是一个巨大的挑战。相比传统方法,基于深度学习的网络通过训练大量数据在识别任务中取得了差异化的效果,但识别效果依旧处于发展过程中。基于此,结合DW卷积和残差连接设计了一种多分支残差模块,该模块通过DW卷积以较小的内存和参数量为代价来加深网络深度,增强网络的特征提取能力;再通过残差连接抑制网络梯度问题和退化问题;另外,提出了一种多分支权重算法,来改善多分支残差模块中各分支的权重分配问题;并将六个以多分支残差模块为主的结构线性连接,组成HCCR识别网络。该模型在CASIA-HWDB1.0、CASIA-HWDB1.1、ICDAR2013数据集上的识别准确率分别达到了97.77%、97.30%、97.64%,表现出高精度的识别效果。展开更多
全球心理健康问题形势严峻,由于心理健康服务的从业人员不足,遭受心理健康困扰的人并不总是能获得专业的心理健康服务.检索式心理健康社区自动问答可以快速地为需要心理健康服务的人提供相应的信息自助服务.与传统检索式社区问答中的文...全球心理健康问题形势严峻,由于心理健康服务的从业人员不足,遭受心理健康困扰的人并不总是能获得专业的心理健康服务.检索式心理健康社区自动问答可以快速地为需要心理健康服务的人提供相应的信息自助服务.与传统检索式社区问答中的文本匹配不同,在匹配支持帖和求助帖时,需要考虑2种不同层面的匹配准则:语义层面和心理层面.为了解决该问题,提出融合角色心理画像的2阶段文本匹配模型(two-stage text matching model integrating characters’mental portrait,T2CMP),该模型引入心理特征用于构建角色心理画像,从而辅助模型理解文本心理层面的内容和匹配关系.同时为了提升检索效率以及减少大量负样例带来的噪声问题,将文本匹配任务拆分为2阶段的序列型子任务.首先针对每条求助帖,使用基于语义的筛选模型甄别出候选支持帖;然后依据用户的角色心理画像,使用多层注意力机制将其与语义信息有效融合,提高模型的总体效果.在MHCQA数据集上的实验结果显示,T2CMP比现有优秀算法拥有更高的F1值.展开更多
高温燃气流风洞的加热段喷注面板由数百个气液同轴离心喷嘴组成,各喷嘴间存在强烈的喷雾干涉现象,导致喷雾场相互耦合。为探究气体中心式气液同轴离心喷嘴喷雾的耦合对雾化特性及流场均匀性的影响,通过实验和仿真的方式研究了不同气液...高温燃气流风洞的加热段喷注面板由数百个气液同轴离心喷嘴组成,各喷嘴间存在强烈的喷雾干涉现象,导致喷雾场相互耦合。为探究气体中心式气液同轴离心喷嘴喷雾的耦合对雾化特性及流场均匀性的影响,通过实验和仿真的方式研究了不同气液比对多喷嘴雾化特性的影响,以及喷嘴间距和喷嘴数目对喷雾流强分布的影响。设计安装多喷嘴的喷注器,搭建喷雾检测实验台,采用高速相机拍摄喷雾图像,采用马尔文激光粒度仪测量喷雾场中的液滴尺寸;并设计了流强测量系统,以测量喷雾场的流强分布。采用流体体积法(Volume of Fluid,VOF)和网格自适应技术(Adaptive Mesh Refinement,AMR)对多喷嘴的耦合喷雾场进行模拟。结果表明,仿真结果与实验测得的流量分布基本吻合;在液体流量较大的工况下,喷雾锥角基本稳定,粒径大小受液膜撞击破碎和液滴撞击聚合双重作用的影响;随着喷嘴间距的增加,喷雾分布的不均匀性增强;并且当存在3个及以上喷嘴时,喷雾场两两相互干涉,在喷雾耦合区域出现流强高峰。展开更多
针对海运货物邮件实体识别中存在识别精度不高、实体边界确定困难的问题,提出一种结合深度学习与规则匹配的识别方法。其中:深度学习方法是在BiLSTM-CRF(Bidirectional Long Short Term Memory-Conditional Random Field)模型的基础上...针对海运货物邮件实体识别中存在识别精度不高、实体边界确定困难的问题,提出一种结合深度学习与规则匹配的识别方法。其中:深度学习方法是在BiLSTM-CRF(Bidirectional Long Short Term Memory-Conditional Random Field)模型的基础上添加词的字符级特征,并融入多头注意力机制以捕获邮件文本中长距离依赖;规则匹配方法则根据领域实体特点制定规则来完成识别。根据货物邮件特点将语料进行标注并划分为:货物名称、货物重量、装卸港口、受载期和佣金五个类别。在自建语料中设置多组对比实验,实验表明所提方法在海运货物邮件实体识别的F1值达到79.3%。展开更多
文摘The Chinese character library is one of the important data structures in the Chinese information Processing system.The behavior of the whole system depends directly on the reasonableness of design for its structure.This paper expounds the structures of RAM-based Chinese character libraries,static and dynamic ,The paper offers a descriptive method for this behavior and inquires into some algorithms related to the structures mentioned above.
文摘The segmentation of individual words into characters is a vital process in handwritten character recognition systems. In this paper, a novel approach is proposed to segment handwritten Arabic text (words). We consider the “Naskh” font style. The segmentation algorithm employs seven agents in order to detect regions where segmentation is illegal. Feature points (end points) are extracted from the remaining regions of the word-image. Initially, the middle of every two successive end points is considered as a candidate segmentation point based on a set of rules. The experimental results are very promising as we achieved a success rate of 86%.
基金This work was supported by the National 863 Program of China(2001AA241251).
文摘A new millet (Setaria italica Beauv) variety, super early-mature millet No.1, was bred by means of gene bank breedingmethod of target characters. This variety has the following outstanding characters. (1) Super early-mature. This varietyonly needs 1550C effective accumulated temperature and can normally maturate in the Bashang Region in Hebei Provinceof Chi na, which can break through the limit zone of millet cultivation and move the cultivation zone northward greatly. (2)Multi-spikes, in addition to the effect tilling at the top, the nodes in the low-middle part also can produce spikes. (3) Sweetstem have high sugar content. The contents of whole-sugar, soluable sugar and deoxidized sugar are 74.8, 200.5, 237.2%higher than the regular varieties respectively. (4) High gross protein content. The content of gross protein is higher thanthe regular varieties by 3.9-30.4%. (5)Changeable grain color. The grain color of super early-mature millet No.1 is red inShijiazhuang, but yellow in the Bashang region. In addition, this variety is characterized by good quality, high yield, andgood synthetic traits
文摘脱机手写中文字符识别(handwritten Chinese character recognition,HCCR)在计算机视觉领域一直是一个巨大的挑战。相比传统方法,基于深度学习的网络通过训练大量数据在识别任务中取得了差异化的效果,但识别效果依旧处于发展过程中。基于此,结合DW卷积和残差连接设计了一种多分支残差模块,该模块通过DW卷积以较小的内存和参数量为代价来加深网络深度,增强网络的特征提取能力;再通过残差连接抑制网络梯度问题和退化问题;另外,提出了一种多分支权重算法,来改善多分支残差模块中各分支的权重分配问题;并将六个以多分支残差模块为主的结构线性连接,组成HCCR识别网络。该模型在CASIA-HWDB1.0、CASIA-HWDB1.1、ICDAR2013数据集上的识别准确率分别达到了97.77%、97.30%、97.64%,表现出高精度的识别效果。
文摘全球心理健康问题形势严峻,由于心理健康服务的从业人员不足,遭受心理健康困扰的人并不总是能获得专业的心理健康服务.检索式心理健康社区自动问答可以快速地为需要心理健康服务的人提供相应的信息自助服务.与传统检索式社区问答中的文本匹配不同,在匹配支持帖和求助帖时,需要考虑2种不同层面的匹配准则:语义层面和心理层面.为了解决该问题,提出融合角色心理画像的2阶段文本匹配模型(two-stage text matching model integrating characters’mental portrait,T2CMP),该模型引入心理特征用于构建角色心理画像,从而辅助模型理解文本心理层面的内容和匹配关系.同时为了提升检索效率以及减少大量负样例带来的噪声问题,将文本匹配任务拆分为2阶段的序列型子任务.首先针对每条求助帖,使用基于语义的筛选模型甄别出候选支持帖;然后依据用户的角色心理画像,使用多层注意力机制将其与语义信息有效融合,提高模型的总体效果.在MHCQA数据集上的实验结果显示,T2CMP比现有优秀算法拥有更高的F1值.
文摘高温燃气流风洞的加热段喷注面板由数百个气液同轴离心喷嘴组成,各喷嘴间存在强烈的喷雾干涉现象,导致喷雾场相互耦合。为探究气体中心式气液同轴离心喷嘴喷雾的耦合对雾化特性及流场均匀性的影响,通过实验和仿真的方式研究了不同气液比对多喷嘴雾化特性的影响,以及喷嘴间距和喷嘴数目对喷雾流强分布的影响。设计安装多喷嘴的喷注器,搭建喷雾检测实验台,采用高速相机拍摄喷雾图像,采用马尔文激光粒度仪测量喷雾场中的液滴尺寸;并设计了流强测量系统,以测量喷雾场的流强分布。采用流体体积法(Volume of Fluid,VOF)和网格自适应技术(Adaptive Mesh Refinement,AMR)对多喷嘴的耦合喷雾场进行模拟。结果表明,仿真结果与实验测得的流量分布基本吻合;在液体流量较大的工况下,喷雾锥角基本稳定,粒径大小受液膜撞击破碎和液滴撞击聚合双重作用的影响;随着喷嘴间距的增加,喷雾分布的不均匀性增强;并且当存在3个及以上喷嘴时,喷雾场两两相互干涉,在喷雾耦合区域出现流强高峰。
文摘针对海运货物邮件实体识别中存在识别精度不高、实体边界确定困难的问题,提出一种结合深度学习与规则匹配的识别方法。其中:深度学习方法是在BiLSTM-CRF(Bidirectional Long Short Term Memory-Conditional Random Field)模型的基础上添加词的字符级特征,并融入多头注意力机制以捕获邮件文本中长距离依赖;规则匹配方法则根据领域实体特点制定规则来完成识别。根据货物邮件特点将语料进行标注并划分为:货物名称、货物重量、装卸港口、受载期和佣金五个类别。在自建语料中设置多组对比实验,实验表明所提方法在海运货物邮件实体识别的F1值达到79.3%。
