ID 原文 译文
25395 针对经典证据理论处理高冲突情况时结果违反常理的问题，在对其原因及各类改进方法进行研究分析的基础上，提出基于冲突关系网络的证据合成方法。 Since the result of classic evidence theory in dealing with high conflict situation is unreasonable, after analyzing the reasons and various improved methods, the evidence combination method based on conflict relation network is proposed．
25396 该方法从整体着眼，充分考虑直接与间接冲突关系，通过构建网络，利用 PageＲank 算法的思想获取证据权值并进行加权修正，最后利用 Dempster 组合规则进行融合，既保留了经典理论优良的数学性质，又较好解决了前述问题。 The method considers direct and indirect relations adequately, takes advantage of the idea of PageＲank algorithm to obtain the weight of evidence by building network. Then the evidence is revised and weighted. Finally，the fusion is accomplished by adopting the combination rule of Dempster. It solves the problem above preferably while keeping the good mathematical properties of evidence theory．
25397 算例分析结果表明，该方法能够客观反映冲突证据所携带的不确定信息，具备一定的优越性。 Analysis of examples indicates that, the proposed method can objectively reflect uncertain information with conflicting evidences well．
25398 目前基于视图的三维模型检索已经成为一个研究热点。 Ｒecently, 3D model retrieval based on views has become a research hot spot．
25399 该方法首先将三维模型表示为二维视图的集合，然后采用深度学习技术进行分类和检索。 In this method, 3D models are represented as a collection of 2D views, which allows deep learning techniques to be used for 3D model classification and retrieval．
25400 但是现有的方法在精度和效率方面都有待提升。 However, current methods need improvements on both accuracy and efficiency．
25401 本文提出了一种新的三维模型检索方法，该方法包括索引建立和模型检索。 We propose a 3D model retrieval method, which includes index building and model retrieval．
25402 在索引建立阶段，选择代表性视图输入到训练好的卷积神经网络(Convolutional Neural Network，CNN)中以提取特征和分类，进而根据特征类别对特征进行组织以建立索引在检索阶段，使用 CNN 和投票算法将输入模型的代表性视图分类为一个类别，然后仅选择这个类别的特征而不是所有类别的特征进行相似度匹配，因此减少了搜索空间。 In the index building stage, representative views are selected and input into a well-learned Convolutional Neural Network (CNN) for feature extraction and classification. Next, the features are organized according to their labels to build indexes. In the retrieval stage, the representative views of the input model are classified into a category with the CNN and voting algorithm, and then only the features of one category rather than all categories are chosen to perform similarity matching. In this way, the searching space for retrieval is reduced．
25403 此外，随着用于检索的视图数量的逐渐增加，一旦可以确定三维模型，检索过程将提前终止。 In addition, the number of the used views for retrieval is gradually increased. Once there is enough evidence to determine a 3D model, the retrieval process will be terminated ahead of time．
25404 实验的数据选用刚 性 三 维 模 型 数 据 集 ModelNet10，ModelNet40 和 非 刚 性 三 维 模 型 数 据 集McGill10。结果表明，该方法在提升检索效率的同时，确保检索准确率分别高达 94% 、92% 和 100% 。 Experiments on the rigid 3D model datasets ModelNet10, ModelNet40, and the non-rigid 3D model dataset McGill10 show that the proposed method can improve the retrieval efficiency substantially while keeping high retrieval accuracy rates at 94% , 92% and 100% , respectively．