共查询到20条相似文献，搜索用时 296 毫秒
1.
随着网络的迅速发展和应用,网络语义标签已经开始广泛地用于图像内容的标注和分享。由于图像本身及不同主体对图像的不同理解会造成图像分析理解的差异,因此如何充分利用网络标签对图像进行准确分析理解成为本文主要研究内容。图像特征包含了图像本身的大量内容信息,为了能够建立图像内容信息与网络标签之间的关系,本文主要工作包括：1,建立低层特征与图像间相似性关系;2,建立基于随机漫步模型平衡图像内容及网络标签间的关系以达到准确对图像内容分析和理解效果。实验结果表明本文所提出方法的有效性和可行性。 相似文献
2.
Automatic web image annotation is a practical and effective way for both web image retrieval and image understanding. However, current annotation techniques make no further investigation of the statement-level syntactic correlation among the annotated words, therefore making it very difficult to render natural language interpretation for images such as ＂pandas eat bamboo＂. In this paper, we propose an approach to interpret image semantics through mining the visible and textual information hidden in images. This approach mainly consists of two parts： first the annotated words of target images are ranked according to two factors, namely the visual correlation and the pairwise co-occurrence; then the statement-level syntactic correlation among annotated words is explored and natural language interpretation for the target image is obtained. Experiments conducted on real-world web images show the effectiveness of the proposed approach. 相似文献
3.
In this paper, we utilize the framework of multi-label learning for face demographic classification. We also attempt to explore
the suitable classifiers and features for face demographic classification. Three most popular demographic information, gender,
ethnicity and age are considered in experiments. Based on the results from demographic classification, we utilize statistic
analysis to explore the correlation among various face demographic information. Through the analysis, we draw several conclusions
on the correlation and interaction among these high-level face semantic, and the obtained results can be helpful in automatic
face semantic annotation and other face analysis tasks. 相似文献
4.
Understanding visual-auditory correlation from heterogeneous features for cross-media retrieval 总被引：1，自引：0，他引：1
Cross-media retrieval is an interesting research topic, which seeks to remove the barriers among different modalities. To enable cross-media retrieval, it is needed to find the correlation measures between heterogeneous low-level features and to judge the semantic similarity. This paper presents a novel approach to learn cross-media correlation between visual features and auditory features for image-audio retrieval. A semi-supervised correlation preserving mapping （SSCPM） method is described to construct the isomorphic SSCPM subspace where canonical correlations between the original visual and auditory features are further preserved. Subspace optimization algorithm is proposed to improve the local image cluster and audio cluster quality in an interactive way. A unique relevance feedback strategy is developed to update the knowledge of cross-media correlation by learning from user behaviors, so retrieval performance is enhanced in a progressive manner. Experimental results show that the performance of our approach is effective. 相似文献
5.
一种小波域K-Means遥感图像分类标注算法 总被引：1，自引：0，他引：1
由于合成孔径雷达图像（遥感）的相干斑噪声数据丰富,导致传统的遥感图像分割方法分割效果不佳,采用学习理论和神经网络改善图像处理性能。根据图像统计特征,采取神经网络语义提出一种高效的图像纹理特征分割方法。首先,利用K-means聚类提取遥感图像的纹理特征,然后根据遥感图像在小波域中的分布特征对其进行滤波,最后利用语义对滤波后的遥感图像纹理特征和灰度组成的矢量进行分割归类,在遥感图像分割中快速标注分类以便于视觉分析。利用区域一致性分割分类,由聚类样本特征匹配进行图像分类标注,对变化检测进行统计分析,过分割或欠分割误差聚类样本不做标注,选取最佳样本聚类k值标注分类结果。 相似文献
6.
7.
旨在介绍一种用于汉英应用翻译的以信息块为视角的语篇信息结构调整方法,这种方法以译语语篇规范为参照,通过语义群归类、信息功能、语篇结构三个层面对原文信息块进行调整、重组,使译文在信息结构上符合译语语篇规范。并通过翻译实例论证这一方法的实用性。 相似文献
8.
Correlation analysis-based image segmentation approach for automatic agriculture vehicle 总被引：1，自引：0，他引：1
INTRODUCTION In recent years,harvesting crop by combine has become more and more popular in China.In2003,combines were used to harvest23%of rice and71%of wheat.The work described here is part of a pro-ject aimed at researching automatic combines who follow the cut-uncut edge of the crops automatically to reduce the driver’s fatigue from the intensive work.Much attention has been focused on vision-bas-ed automatic agriculture vehicle.Image segmenta-tion is the key to extracting the guida… 相似文献
9.
刘锦凤 《读与写:教育教学刊》2009,6(7):9-10
The past several decades have witnessed a phenomenal growth in interest in text analysis, in which different kinds of approaches have been studied and applied in this field. This paper aims at analyzing the introduction of a cho-sen CNN News from a functional linguis.tic approach, which is mainly realized through cohesive means and textual infor-mation. The study shows that in written text, well-organized semantic cohesive means and textual information are of great significance for readers to follow the movement of an idea from one sentence to another. Therefore, functional approach plays a momentous role in the analysis of a text. 相似文献
10.
线性构造是指遥感图像上直线状或者曲线状的地表线性影像,为了对线性构造目视解译和计算机自动解译结果进行比较,基于高分一号卫星遥感影像进行 RGB 彩色合成和低通滤波处理,达到图像增强的目的,并分别采用人工目视解译和计算机自动解译方法提取线性构造信息,对得到的数据进行统计分析,分别作相应的雷达图和密度图,并进行非参数检验和相关性分析。结果显示,雷达图所体现的线性构造的最优选方位均为50°;密度图高值区均呈现 NE-SW 两向延长分布;非参数检验的显著性值为 0.843;Pearson 相关系数为 0.552。研究表明,人工目视解译和计算机自动解译结果一致性较好,并具有较强的内在相关性,因此在线性构造研究中,计算机自动解译可作为对传统人工目视解译的一种辅助和验证手段。 相似文献
11.
针对图像底层视觉特征与高层情感语义间的语义鸿沟,分析色彩与情感的关系,建立情感空间,采用概率神经网络建立图像颜色特征与情感空间的映射,融合图像底层视觉特征和人类情感特征,提出并实现了一种感性图像检索方法。实验结果证明了该方法具有较好的效果,为图像情感语义的研究提供了一种行之有效的解决方案。 相似文献
12.
肖星星 《内江师范学院学报》2012,27(11):100-102
文本分析方法是宪法解释方法中比较常用的方法之一。宪法文本分析,是指以对宪法条文的文字和词组为基础,结合宪法条文的上下文,对宪法条文的字义作极其狭隘的、准确的解释。本文从宪法文本分析的概念和特点、宪法文本的价值、如何进行宪法文本分析三个方面作了必要的分析。 相似文献
13.
14.
盛仲飙 《渭南师范学院学报》2014,(3):43-46
基于互信息的图像检索方法是根据图像的可视性及相似性测度对图像进行检索的方法．首先利用互信息生成一种在统计上有代表性的视觉模式,使用这种模式的分布作为图像内容的描述符;接着利用该内容的描述,设计其互信息的计算方法以衡量图像的相似性．实验结果表明,在图像检索中,相对于其他如KL散度和也规范等方法,互信息是一种更为有效的衡量相似性的方法． 相似文献
15.
Yun-Ping Ge Len Unsworth Kuo-Hua Wang Huey-Por Chang 《Research in Science Education》2018,48(6):1409-1431
From a social semiotic perspective, image designs in science textbooks are inevitably influenced by the sociocultural context in which the books are produced. The learning environments of Australia and Taiwan vary greatly. Drawing on social semiotics and cognitive science, this study compares classificational images in Australian and Taiwanese junior high school science textbooks. Classificational images are important kinds of images, which can represent taxonomic relations among objects as reported by Kress and van Leeuwen (Reading images: the grammar of visual design, 2006). An analysis of the images from sample chapters in Australian and Taiwanese high school science textbooks showed that the majority of the Taiwanese images are covert taxonomies, which represent hierarchical relations implicitly. In contrast, Australian classificational images included diversified designs, but particularly types with a tree structure which depicted overt taxonomies, explicitly representing hierarchical super-ordinate and subordinate relations. Many of the Taiwanese images are reminiscent of the specimen images in eighteenth century science texts representing “what truly is”, while more Australian images emphasize structural objectivity. Moreover, Australian images support cognitive functions which facilitate reading comprehension. The relationships between image designs and learning environments are discussed and implications for textbook research and design are addressed. 相似文献
16.
葡萄与葡萄酒质量的综合评价 总被引：1，自引：0，他引：1
从体现酒品质量的感官评价和酿酒葡萄与葡萄酒的理化指标两方面入手,运用t检验、主成分分析法、模糊c均值聚类分析、多元线性回归分别建立显著性差异检验、主成分、聚类分析、质量影响和论证等模型,并运用SPSS、Matlab7软件,得出两组评酒员的评价结果差异性和可信度,酿酒葡萄的分级,酿酒葡萄与葡萄酒的理化指标之间的联系,最后分析并论证了葡萄、葡萄酒的理化指标与葡萄酒质量的关系. 相似文献
17.
Eric P. Kumpf 《Technical Communication Quarterly》2013,22(4):401-424
Visual metadiscourse can provide design criteria for authors when considering the needs and expectations of readers. The linguistic concept of metadiscourse is expanded from the textual realm to the visual realm, where authors have many necessary design considerations as they attempt to help readers navigate through and understand documents. These considerations, both textual and visual, also help construct the ethos of authors, as design features reveal awareness of visual literacy and of the communication context. Visual metadiscourse complements textual metadiscourse in emphasizing the necessity of rhetoric in technical communication. 相似文献
18.
The article argues for the importance of studying word‐recognition strategies in the assessment of dyslexia. The consensus model of word recognition in cognitive neuropsychology, the dual‐route model, is critically discussed in the light of attacks from computational models in modern connectionism. Despite the criticism, it is argued that a dual‐route model can serve important functions in the context of dyslexia assessment. We propose a specific version of the dual‐route model from which we construct a series of tasks for diagnosing subprocesses in word recognition. A computer‐based test battery is described, involving oral reading tasks, lexical decision, rhyme detection, letter categorization, visual analysis, naming, and semantic categorization. Latencies and accuracy scores are recorded. The usefulness of this assessment approach is demonstrated in two case studies of 15‐year‐old dyslexic boys. 相似文献
19.
文本语言分析是对文学文本进行的理论批评。在当今文化分析方兴未艾的理论背景下,反思新批评以来的文本语言分析理论的得失,完成文本分析到文化分析的过渡,理应成为当务之急。新批评的文本分析理论、结构主义的文本分析理论和解构主义的文本分析理论,固然重视了对文本的语言组织特点加以分析,但是对文本语言潜在情感因素的重视是远远不够的,对此提出对文本语言情感因素的挖掘,重视对文本语言潜在情感因素的分析。 相似文献
20.
工具痕迹是一种出现在刑事案件中的重要物证,同时也是串并同类案件的重要依据。充分利用案件现场的工具痕迹有着十分重要的意义。根据小波变换原理,从纹理分析的角度出发,对工具痕迹图像进行二维小波变换,提取灰度共生矩阵的相关特征参数构造BP神经网络对痕迹图像进行分类识别,实验表明,该方法能细致的描述工具痕迹图像的纹理特征,可以辅助刑事技术人员进行工具痕迹图像特征的比对分析。 相似文献