共查询到18条相似文献,搜索用时 123 毫秒
1.
2.
3.
4.
5.
6.
7.
知识表示中的不确定性 总被引:33,自引:1,他引:32
知识表示一直是人工智能研究中的一个瓶颈,其难点在于知识中隐含有不确定性,即模糊性和随机性。意提出用云模型3个数字特征(期望值,熵,超熵)来描述一个定性概念,用熵来关联模糊性和随机性。代表定性概念的云的某一次定量值,被称为云滴,要以用它对此概念的贡献度来衡量,许许多多云滴构成云,实现实性和定量之间的随时转换,反映了知识表示中的不确定性。论以此对我国农历24个节气进行了新的最化解释。云方法已经用于数据开采、智能控制、跳频电台和大系统效能评估中,取得明显的效果。 相似文献
8.
9.
针对甲状腺肿瘤超声图像对比度低和SPECT图像边界模糊的特点,结合多尺度几何分析和单尺度稀疏表示的思想,提出了一种 Shearlet 变换与稀疏表示相结合的图像融合算法。首先,用该变换对已经精确配准的源图像进行分解,得到图像的高低频子带系数。对稀疏性较差的低频子带系数进行字典训练并求解其稀疏表示系数,并采用能量值取大的规则进行融合。高频子带系数采用区域拉普拉斯能量和的规则。最后,用 Shearlet 逆变换得到融合图像。实验结果表明,此算法在主观视觉效果和客观评价指标上优于多尺度融合方法和单尺度下基于稀疏表示的图像融合方法。 相似文献
10.
本课题组在开发CIMS特征造型软件过程中试作了CAD与CAPP的直接接口,从中发现CAD与CAPP对于零件的特征建模表示方法有不同的理解和要求。本文列举这些差异,以供CIMS和并行工程环境下建立更完善的集成产品信息模型作参考。 相似文献
11.
12.
Design representation of parts during conceptual design stage has been a challenging task because of the incompleteness of the information available. Traditional geometric design requires too much information on the geometric attributes, and does not consider the functionality of the part any more than that provided by individual features. Functional design in the form of verb, noun representation does not have adequate correlation with the geometric design. In this article, we propose a new form of the design representation of parts during conceptual design. The representation is calledsketching abstraction. In this representation, the discretionary geometry of the part that has functional relevance is captured using functional features, while the non-discretionary geometry is represented using a linkage mechanism. The functional features are related to the part function using data structures calledfunction-form matrices. The sketching abstraction is annotated with a set of primitives, and a grammar has been developed that parses the sketch to extract a set of canonical relationships between the functional features. These relationships can be used to extract part designs that are functionally similar but geometrically dissimilar. The sketching abstraction has a relationship with the solid model of the part as well. Thus we attempt to bridge the gap between function and form representations and provide the designer with a tool that can be used for generating design alternatives. We illustrate the theory developed in the domain of stamped metal parts. 相似文献
13.
In a multidisciplinary design environment, such as the architecture, engineering and construction (AEC) domain, the various designers will have their own concepts and representations of the design object making communication in such an environment a complex task. Computer-aided design (CAD) modelling is seen as the vehicle for interdisciplinary communication and integration. However, this paper stresses the need for a multiple view approach based upon an assumption that different views of an object are based on different functional contexts. The paper demonstrates that an understanding of concepts such as function, purpose and intent is critical since the representation of the functional properties of design objects is the underlying basis for the formation of different representations and the coordination of these representations. The paper puts forward definitions for function and purpose which allow for the representation of these properties of a design object and the communication of intent and effects. 相似文献
14.
15.
The conventional sparse representation-based image classification usually codes the samples independently, which will ignore the correlation information existed in the data. Hence, if we can explore the correlation information hidden in the data, the classification result will be improved significantly. To this end, in this paper, a novel weighted supervised spare coding method is proposed to address the image classification problem. The proposed method firstly explores the structural information sufficiently hidden in the data based on the low rank representation. And then, it introduced the extracted structural information to a novel weighted sparse representation model to code the samples in a supervised way. Experimental results show that the proposed method is superiority to many conventional image classification methods. 相似文献
16.
研究了专家系统中知识的多重表示情况下推理的非单调真值维护。指出了专家系统中知识的多重表示的必要性,以及知识的多重表示情况下可能出现的推理的不一致性。着重讨论了知识的层次表示和优先级的划分。基于领域知识的优先级,分层ATMS被用来实现推理的非单调真值维护。给出了Stratified ATMS的推理和真值维护的算法。 相似文献
17.
目的 为了有效去除图像中的椒盐噪声,提高图像质量。方法 文中将可分离字典和低秩表示结合,提出基于可分离字典的稀疏和低秩表示算法(SLRR–SD)。首先,使用可分离字典代替传统的过完备字典可分离字典可以对二维图像直接表示。其次,使用Frobenius范数对分离字典进行约束以挖掘字典内部的低秩性。此外,为了挖掘图像内部的稀疏结构,对表示系数使用稀疏约束进一步提升表示的有效性。结果 提出的算法在噪声强度为5%、10%、20%和30%下,PSNR/FSIM的平均值分别为32.736/0.975、29.769/0.957、29.295/0.951和26.768/0.921。结论 文中算法保留了相邻列之间的相关性,并且可分离字典优化过程也降低了计算负担。实验结果表明,该算法在保留原图像信息的同时能更好地完成去噪任务。 相似文献