RAFNet: RGB-D attention feature fusion network for indoor semantic segmentation

Semantic segmentation based on the complementary information from RGB and depth images has recently gained great popularity, but due to the difference between RGB and depth maps, how to effectively use RGB-D information is still a problem. In this paper, we propose a novel RGB-D semantic segmentation network named RAFNet, which can selectively gather features from the RGB and depth information. Specifically, we construct an architecture with three parallel branches and propose several complementary attention modules. This structure enables a fusion branch and we add the Bi-directional Multi-step Propagation (BMP) strategy to it, which can not only retain the feature streams of the original RGB and depth branches but also fully utilize the feature flow of the fusion branch. There are three kinds of complementary attention modules that we have constructed. The RGB-D fusion module can effectively extract important features from the RGB and depth branch streams. The refinement module can reduce the loss of semantic information and the context aggregation module can help propagate and integrate information better. We train and evaluate our model on NYUDv2 and SUN-RGBD datasets, and prove that our model achieves state-of-the-art performances.     

基于描述逻辑的OntoUML模型的形式化表示

将形式本体理论应用于UML语言的改进是UML的形式化研究的主要方法之一,其中On-toUML是一个概念上完整、语义上清晰和更具本体意义的UML版本。虽然OntoUML基于本体对UML进行了扩充,提供了更为丰富,更能表达现实世界语义的建模原语,但其表达方式不利于在信息系统的设计与开发中应用。本文基于描述逻辑对OntoUML中的关键建模原语及其关系给出了SHIQ形式化表示方法,并进行了实例研究。该方法表达精练、语义清晰,不仅能够表达逻辑清晰的模型,而且保证模型能够正确反映领域事实。这种形式化方法为信息系统概念建模方法提供了理论和应用支撑,在一定程度上推动了OntoUML更为广泛的应用。     

一种新颖的低层视觉特征到高层语义的映射方法*

通过在低层视觉特征上提取图像的局部颜色和纹理特征,再利用模糊神经网络建立低层视觉特征和高层语义特征之间的映射,实现了图像属性信息的有效传递和可靠检索.最后,为了提高检索效率,把相关反馈引进到图像检索系统中.实验证明,该方法取得了较好的检索查全率和准确率.     

基于图的语义网络构造算法研究

在自然语言处理的基础上,将中医领域本体与图的基本性质相结合,提出了构造中医诊疗知识语义网络算法。给出了图的基本性质、语义网络的图定义,并为知识的网络表示建立了相应的数据结构。将算法用于中医领域语义网络和语义场构建。试验结果表明,该算法具有较好的实用性。提出的基于图的语义网络构造算法,对于文本挖掘、知识获取技术研究有重要意义,也是中医专家知识获取的有效方法。     

Semantic weak signal tracing

The weak signal concept according to Ansoff has the aim to advance strategic early warning. It enables to predict the appearance of events in advance that are relevant for an organization. An example is to predict the appearance of a new and relevant technology for a research organization. Existing approaches detect weak signals based on an environmental scanning procedure that considers textual information from the internet. This is because about 80% of all data in the internet are textual information. The texts are processed by a specific clustering approach where clusters that represent weak signals are identified. In contrast to these related approaches, we propose a new methodology that investigates a sequence of clusters measured at successive points in time. This enables to trace the development of weak signals over time and thus, it enables to identify relevant weak signal developments for organization’s decision making in strategic early warning environment.     

Text classification using genetic algorithm oriented latent semantic features

In this paper, genetic algorithm oriented latent semantic features (GALSF) are proposed to obtain better representation of documents in text classification. The proposed approach consists of feature selection and feature transformation stages. The first stage is carried out using the state-of-the-art filter-based methods. The second stage employs latent semantic indexing (LSI) empowered by genetic algorithm such that a better projection is attained using appropriate singular vectors, which are not limited to the ones corresponding to the largest singular values, unlike standard LSI approach. In this way, the singular vectors with small singular values may also be used for projection whereas the vectors with large singular values may be eliminated as well to obtain better discrimination. Experimental results demonstrate that GALSF outperforms both LSI and filter-based feature selection methods on benchmark datasets for various feature dimensions.     

局部自适应复合绝缘子检测与憎水性分类网络

憎水性等级(Hydrophobicity Class,HC)是衡量绝缘子性能的重要指标之一。在实际环境的多种因素作用下绝缘子伞裙表面存在局部憎水性差异,为了准确识别绝缘子的性能,本文提出了一种基于深度学习的局部自适应绝缘子检测与憎水性分类模型。首先,通过绝缘子分割模块分离绝缘子与背景区域,为后续针对绝缘子区域的操作提供分割信息;然后将绝缘子区域划分为固定大小的图像块,在缩小分辨率减小运算难度的同时保留了绝缘子表面的细节信息;最后通过憎水性分类模块分析图像块内绝缘子的憎水性。实验使用巡检维护现场的绝缘子图片作为样本集,分阶段构建模型,分别对分割阶段和憎水性分类阶段的准确性进行评估。实验结果显示分割阶段模块能有效识别绝缘子和背景区域,交叉验证的测试集准确率均大于97.21%,并且憎水性分类阶段模块能准确分析绝缘子憎水性,对140幅测试图片的识别准确率达到98.65%。经过实验证明本文提出的模型在复杂自然环境中检测绝缘子性能是一种有效的解决方案。     

多类型分类器融合的文本分类方法研究

传统的文本分类方法大多数使用单一的分类器,而不同的分类器对分类任务的侧重点不同,就使得单一的分类方法有一定的局限性,同时每个特征提取方法对特征词的考虑角度不同。针对以上问题,提出了多类型分类器融合的文本分类方法。该模型使用了word2vec、主成分分析、潜在语义索引以及TFIDF特征提取方法作为多类型分类器融合的特征提取方法。并在多类型分类器加权投票方法中忽略了类别信息的问题,提出了类别加权的分类器权重计算方法。通过实验结果表明,多类型分类器融合方法在二元语料库、多元语料库以及特定语料库上都取得了很好的性能,类别加权的分类器权重计算方法比多类型分类器融合方法在分类性能方面提高了1.19%。     

基于分布的中文词表示研究

针对基于分布的中文词表示构造过程中的参数选择问题进行了系统性的研究。选择了六种参数进行对比实验,在中文语义相似度任务上对不同参数设置下得到的中文词表示的质量进行了评估。实验结果表明,通过选择合适的参数,基于分布的词表示在中文语义相似度任务上能够得到较高的性能,而且,这种高维的词分布表示的质量甚至优于目前流行的基于神经网络（Skip-gram）或矩阵分解（GloVe）得到的低维的词表示。     

基于web的外贸企业资金流管理平台的设计

本文针对外贸型企业对外贸易服务流程中的资金流动和资金流管理中存在的问题,设计了一个基于Web的外贸企业资金流管理平台,实现跨地域、多角色、实时的资金流录入、统计和分析,使企业能够高效准确地管理业务和资金流,能够随时统计资金的流向以及分析盈利情况;着重论述外贸企业资金流管理需求、基于Web的系统框架设计、数据加密存储、加密传输、用户角色及权限控制等。