A key to understand Immanuel Kant's philosophy is his views on humanity,and the cornerstone of Kant’s philosophy is the idea that“beauty is a symbol of morality”.From the perspective of his views on humanity,we...A key to understand Immanuel Kant's philosophy is his views on humanity,and the cornerstone of Kant’s philosophy is the idea that“beauty is a symbol of morality”.From the perspective of his views on humanity,we can have a deep understanding of Kant*s analysis of beauty and the sublime and his aesthetics.The focus of Kant's aesthetics is the realization of human freedom and the sublime,and this is of special humanistic feature.展开更多
The construction of human settlement environment is one of the important contents in the domain of sustainable development. We try to annotate human settlement environment for "village- in-city" from the ang...The construction of human settlement environment is one of the important contents in the domain of sustainable development. We try to annotate human settlement environment for "village- in-city" from the angle of fringe view. According to the macro-system thought of human settlement environment science, the evolvement, which can be generalized into four phases, connotation and problems (including social, economic, and environmental problems) of "village-in-city" are discussed primarily in this paper. Some domestic and international researches and practices are also summarized and assessed in the paper. Based on the analysis, some appropriate clues and suggestions for the construction of "village-in-city" have been put forward.展开更多
作为人工智能计算机视觉领域一项重要的任务,3D人体姿态估计受到了广泛的关注,并成功地应用在人机交互、电影游戏制作等领域。然而,3D人体姿态估计仍然面临着很大的挑战,主要是人体遮挡问题和数据集视角冗余问题,这些问题严重影响了3D...作为人工智能计算机视觉领域一项重要的任务,3D人体姿态估计受到了广泛的关注,并成功地应用在人机交互、电影游戏制作等领域。然而,3D人体姿态估计仍然面临着很大的挑战,主要是人体遮挡问题和数据集视角冗余问题,这些问题严重影响了3D人体姿态估计结果精度与速度的提升。本文提出了一种基于多特征提取的3D人体姿态估计方法。首先通过采集多个相机视角下的图片数据,将所采图片数据放入2D人体关节点检测网络模型中,得到人体2D关节点。接着将采集到的人体数据输入到关节点置信度计算网络模型,得到视角图片中各个关节点的权重值。随后将2D人体关节点热图通过一个热图权重计算网络计算出热图权重,将各个视角下的权重特征计算融合得到加权后的2D人体关节点热图。最后将所得加权后的2D人体关节点热图和视角图片中各个关节点的权重值输入到三角化算法中,映射得到空间中的3D人体关节点。本文的关键思想是设计一个关节点置信度计算网络从输入图像中学习每个关节的置信度权重,同时提取了反映热图特征质量的权重矩阵,以提高遮挡视图中热图的特征质量。此外,使用感知哈希算法对Occlusion-Person数据集进行去视角实验,在保证结果准确性的同时提高了模型推理速度。本文方法是端到端可微的,可以显著地提高算法效率和鲁棒性。本文在Human3.6M和Occlusion-Person两个公共数据集上使用平均关节位置误差(Mean Per Joint Position Error,MPJPE)指标对该方法进行评估,分别取得27.3 mm和9.7 mm的结果。实验结果表明,该算法与最先进的方法相比,性能有了显著提升。展开更多
传统编目分类和规则匹配方法存在工作效能低、过度依赖专家知识、缺乏对古籍文本自身语义的深层次挖掘、编目主题边界模糊、较难实现对古籍文本领域主题的精准推荐等问题。为此,本文结合古籍语料特征探究如何实现精准推荐符合研究者需...传统编目分类和规则匹配方法存在工作效能低、过度依赖专家知识、缺乏对古籍文本自身语义的深层次挖掘、编目主题边界模糊、较难实现对古籍文本领域主题的精准推荐等问题。为此,本文结合古籍语料特征探究如何实现精准推荐符合研究者需求的文本主题内容的方法,以推动数字人文研究的进一步发展。首先,选取本课题组前期标注的古籍语料数据进行主题类别标注和视图分类;其次,构建融合BERT(bidirectional encoder representation from transformers)预训练模型、改进卷积神经网络、循环神经网络和多头注意力机制的语义挖掘模型;最后,融入“主体-关系-客体”多视图的语义增强模型,构建DJ-TextRCNN(DianJi-recurrent convolutional neural networks for text classification)模型实现对典籍文本更细粒度、更深层次、更多维度的语义挖掘。研究结果发现,DJ-TextRCNN模型在不同视图下的古籍主题推荐任务的准确率均为最优。在“主体-关系-客体”视图下,精确率达到88.54%,初步实现了对古籍文本的精准主题推荐,对中华文化深层次、细粒度的语义挖掘具有一定的指导意义。展开更多
基金Key project of Humanities and Social Sciences of Anhui provincial Education Department(SK2017A0380)General project of Humanities and Social Sciences of Anhui provincial Education Department(SKHS2016B08)School-level Research Platform(KYPT201816)
文摘A key to understand Immanuel Kant's philosophy is his views on humanity,and the cornerstone of Kant’s philosophy is the idea that“beauty is a symbol of morality”.From the perspective of his views on humanity,we can have a deep understanding of Kant*s analysis of beauty and the sublime and his aesthetics.The focus of Kant's aesthetics is the realization of human freedom and the sublime,and this is of special humanistic feature.
文摘The construction of human settlement environment is one of the important contents in the domain of sustainable development. We try to annotate human settlement environment for "village- in-city" from the angle of fringe view. According to the macro-system thought of human settlement environment science, the evolvement, which can be generalized into four phases, connotation and problems (including social, economic, and environmental problems) of "village-in-city" are discussed primarily in this paper. Some domestic and international researches and practices are also summarized and assessed in the paper. Based on the analysis, some appropriate clues and suggestions for the construction of "village-in-city" have been put forward.
文摘作为人工智能计算机视觉领域一项重要的任务,3D人体姿态估计受到了广泛的关注,并成功地应用在人机交互、电影游戏制作等领域。然而,3D人体姿态估计仍然面临着很大的挑战,主要是人体遮挡问题和数据集视角冗余问题,这些问题严重影响了3D人体姿态估计结果精度与速度的提升。本文提出了一种基于多特征提取的3D人体姿态估计方法。首先通过采集多个相机视角下的图片数据,将所采图片数据放入2D人体关节点检测网络模型中,得到人体2D关节点。接着将采集到的人体数据输入到关节点置信度计算网络模型,得到视角图片中各个关节点的权重值。随后将2D人体关节点热图通过一个热图权重计算网络计算出热图权重,将各个视角下的权重特征计算融合得到加权后的2D人体关节点热图。最后将所得加权后的2D人体关节点热图和视角图片中各个关节点的权重值输入到三角化算法中,映射得到空间中的3D人体关节点。本文的关键思想是设计一个关节点置信度计算网络从输入图像中学习每个关节的置信度权重,同时提取了反映热图特征质量的权重矩阵,以提高遮挡视图中热图的特征质量。此外,使用感知哈希算法对Occlusion-Person数据集进行去视角实验,在保证结果准确性的同时提高了模型推理速度。本文方法是端到端可微的,可以显著地提高算法效率和鲁棒性。本文在Human3.6M和Occlusion-Person两个公共数据集上使用平均关节位置误差(Mean Per Joint Position Error,MPJPE)指标对该方法进行评估,分别取得27.3 mm和9.7 mm的结果。实验结果表明,该算法与最先进的方法相比,性能有了显著提升。
文摘传统编目分类和规则匹配方法存在工作效能低、过度依赖专家知识、缺乏对古籍文本自身语义的深层次挖掘、编目主题边界模糊、较难实现对古籍文本领域主题的精准推荐等问题。为此,本文结合古籍语料特征探究如何实现精准推荐符合研究者需求的文本主题内容的方法,以推动数字人文研究的进一步发展。首先,选取本课题组前期标注的古籍语料数据进行主题类别标注和视图分类;其次,构建融合BERT(bidirectional encoder representation from transformers)预训练模型、改进卷积神经网络、循环神经网络和多头注意力机制的语义挖掘模型;最后,融入“主体-关系-客体”多视图的语义增强模型,构建DJ-TextRCNN(DianJi-recurrent convolutional neural networks for text classification)模型实现对典籍文本更细粒度、更深层次、更多维度的语义挖掘。研究结果发现,DJ-TextRCNN模型在不同视图下的古籍主题推荐任务的准确率均为最优。在“主体-关系-客体”视图下,精确率达到88.54%,初步实现了对古籍文本的精准主题推荐,对中华文化深层次、细粒度的语义挖掘具有一定的指导意义。
