基于稀疏学习的人脸语义子空间提取

人脸语义检索在识别技术中有着重要的作用,如表情检索、性别判断、年龄估计等,上述识别技术通过捉捕人脸语义信息来实现。研究将人脸语义信息融入到人脸检索中,提出一种基于稀疏学习的人脸语义子空间提取方法。语义子空间学习被分为字典构建和稀疏学习2个部分。在字典构建的过程中,给出语义差的方法来对互斥语义进行计算,使提取的某类语义不受其他类语义干扰语义子空间,并对不同语义环境和不同语义差组合进行测试。在稀疏学习部分,使用Lasso算法对其进行改进。实验结果表明,与传统Fisher方法相比,该方法撇除其他语义干扰的子空间稳定性更强,且有一定的降维效果。     

基于张量子空间学习的人行为识别方法

提出了一种基于张量子空间学习降维人体高维侧影数据的人行为识别方法。给定一个动作的人侧影图像序列,首先用张量子空间学习方法将目标高维侧影图像投影到低维子空间来描述人运动的时空特性,并同时尽可能地保持目标侧影图像中像素之间的空间几何信息,然后用Hausdorff距离度量动作之间的相似性,并在最近邻距离框架下对动作进行分类识别。为验证本文算法的有效性,设计了动作识别和鲁棒性测试2个实验。实验结果表明提出的算法不仅能够有效地对人行为进行识别,且具有较强的鲁棒性。     

基于改进的数据驱动决策树分析的3维人体运动检索

随着大量3维人体运动捕获数据库的出现,使得如何对人体运动数据进行高效分析和处理,从而有效利用运动捕获数据库成为一个新的挑战。为了高效地进行3维人体运动检索,首先通过从人体运动中提取一种基于3维空间变换特征规律的空间变换特征和运动的一些关键的时间特性来得到人体运动的3维时序特征;然后针对不同的训练需求,通过改进的数据驱动决策树的学习方法来分析关节点对运动相似的不同影响,并在检索过程中按照不同影响程度依次对关键点进行相似度计算;最终实现了一个高效的运动检索仿真系统。     

基于CNN的改进串行混合网络及多传感器信息融合的运动意图识别研究

为了提高基于传感器信号的运动意图识别准确度,开展了基于深度学习和多传感器信息融合的运动意图识别方法的研究.首先,在平地、10°上/下坡、上/下楼梯,共5种步态模式下,提取了80名受试者的运动传感器数据,采用非线性降维Diffusion Maps法(DM)对数据进行降维,利用短期傅里叶变化(SIFT)提取数据特征;构建基于卷积神经网络(CNN)和埃尔曼神经网络(ENN)的新型串行混合网络模型(CNN-ENN),并完成了模型的训练;随机选取了5名志愿者参与实验测试.实验结果表明CNN-ENN改进串行混合网络模型对多传感器融合信息的平均识别准确度为95.86％,分别比传统的运动意图识别算法CNN和基于集成学习的加权投票算法(MFR)高10.21％和16.37％.实践表明深度学习模型与多传感器信息融合技术相结合的方法对运动意图识别精度的提高有很好的指导价值.     

面向跨被试RSVP的多特征低维子空间嵌入的ERP检测

基于快速序列视觉呈现（rapid serial visual presentation, RSVP）范式的目标图像检索借助于人脑在看到目标图像时产生的事件相关电位（event-related potentials, ERP）来完成复杂目标图像检索。在应用RSVP范式进行复杂目标图像检索时存在跨时段甚至跨被试的问题。对此，本文提出了一种面向跨被试RSVP的多特征低维子空间嵌入的ERP检测方法，首先采用迁移学习方法中的欧式空间对齐对不同被试的数据进行对齐，其次将来自不同空间的特征分别进行有监督降维、重构。最终采用留一被试法作为检验方法、平衡准确率作为评价指标，在PhysioNetRSVP数据集以及清华RSVP数据集下共计14个长度分段中，有12个长度分段达到最优分类结果。结果表明本文提出的多特征低维子空间嵌入方法能够有效提升ERP检测时的稳定性。     

加工面点云数据深度学习的加工特征自动识别

加工特征识别是实现CAD/CAPP/CAM系统集成的关键技术.针对传统基于符号推理加工特征识别模式存在鲁棒性问题,提出一种基于加工面点云数据深度学习的加工特征自动识别方法;基于PointNet点云识别框架,构建了一个面向加工面点云数据的加工特征自动识别卷积神经网络;通过收集CAD模型中的加工特征面集和采样点云,构建了适...     

对步态空时数据的连续特征子空间分析

提出一种基于空时特征提取的人体步态识别算法。连续的特征子空间学习依次提取出步态的时间与空间特征：第一次特征子空间学习对步态的频域数据进行主成分分析,步态数据被转化为周期特征矢量;第二次特征子空间学习对步态数据的周期特征矢量形式进行主成分分析加线性判别分析的联合分析,步态数据被进一步转化为步态特征矢量。步态特征矢量同时包含运动的周期特征以及人体的形态特征,具有很强的识别能力。在USF步态数据库上的实验结果显示,该算法识别率较其他同类算法有明显提升。     

基于双向参考索引的大规模人体运动数据库的检索

因为运动特征数据的高维复杂性,采用非线性的Isomap流形学习的降维算法来对运动特征数据进行降维,为了能让Isomap方法处理训练数据集之外的数据,通过学习主成分特征核函数逼近降维结果,以扩展传统Isomap的局限性.在运动数据降维之后,为大规模运动捕获数据库建立一种双向参考索引(DRI),在检索过程中索引用来排除绝大部分与查询例子无关的运动数据,这样运动检索中的相似度的计算通过索引被缩小到一个小范围候选数据集合中,避免了大量不必要的匹配开销,从而提高了检索的效率.     

基于人体姿势编码的运动数据检索

针对大规模运动捕获数据的检索问题,提出一种基于人体姿势编码的相似运动检索方法.首先将形成人体姿势的9段主要骨骼的运动空间划分成多个子空间,以各骨骼所处的子空间位置作为人体姿势的特征表示;然后根据特征表示对运动数据进行逐帧编码,将编码相同的连续帧划分成等价段;最后建立以等价段为基础的倒排索引来检索候选运动,并通过编码匹配...     

基于分块主成分分析的人体运动合成

在高维运动数据处理中,传统降维方法过度关注于保护拓扑结构和重构信息,导致低维子空间参数具有不可理解性。针对该问题,提出一种基于分块主成分分析的运动合成方法。对根据人体骨骼结构特点分组的运动数据进行分块主成分分析,获得具有特定语义的低维子空间参数。以跳跃运动为例进行实验,结果表明,该方法通过直观地改变具有语义的运动参数,可实时合成满足要求的运动结果。     

基于集成多示例学习决策树分析的三维运动检索

随着运动捕获设备的普及,大量的运动数据可以直接得到,从而使得大规模的运动数据库的建立成为可能.在此背景下,研究以检索为核心的运动捕获数据处理技术就显得十分重要了.提出了一种对运动捕获数据中的人体的各个关节点提取一种基于三维空间变换规律的三维空间特征的方法,并基于空间运动连续性引入了关键空间的概念.针对各关节点空间特征相对保持独立的特性,用每个关节点作为索引,并通过集成学习的多示例决策树的学习方法去分析关节点对运动相似的不同影响,最终实现了一个高效的运动检索系统.     

Human motion retrieval using topic model

Content‐based human motion retrieval is important for animators with the development of motion editing and synthesis, which need to search similar motions in large databases. Obtaining text‐based representation from quantization of mocap data turned out to be efficient. It becomes a fundamental step of many researches in human motion analysis. Geometric features are one of these techniques, which involve much prior knowledge and reduce data redundancy of numerical data. We describe geometric features as basic unit to define human motions (also called mo‐words) and view a human motion as a generative process. Therefore, we obtain topic motions, which possess more semantic information using latent Dirichlet allocation by learning from massive training examples in order to understand motions better. We combine probabilistic model with human motion retrieval and come up with a new representation of human motions and a new retrieval framework. Our experiments demonstrate its advantages, both for understanding motions and retrieval. Copyright © 2011 John Wiley & Sons, Ltd.     

A survey on ensemble learning

Despite significant successes achieved in knowledge discovery,traditional machine learning methods may fail to obtain satisfactory performances when dealing with complex data,such as imbalanced,high-dimensional,noisy data,etc.The reason behind is that it is difficult for these methods to capture multiple characteristics and underlying structure of data.In this context,it becomes an important topic in the data mining field that how to effectively construct an efficient knowledge discovery and mining model.Ensemble learning,as one research hot spot,aims to integrate data fusion,data modeling,and data mining into a unified framework.Specifically,ensemble learning firstly extracts a set of features with a variety of transformations.Based on these learned features,multiple learning algorithms are utilized to produce weak predictive results.Finally,ensemble learning fuses the informative knowledge from the above results obtained to achieve knowledge discovery and better predictive performance via voting schemes in an adaptive way.In this paper,we review the research progress of the mainstream approaches of ensemble learning and classify them based on different characteristics.In addition,we present challenges and possible research directions for each mainstream approach of ensemble learning,and we also give an extra introduction for the combination of ensemble learning with other machine learning hot spots such as deep learning,reinforcement learning,etc.     

A novel block intensity comparison code for video classification and retrieval

Appropriate organization of video databases is essential for pertinent indexing and retrieval of visual information. This paper proposes a new feature called block intensity comparison code (BICC) for video classification and retrieval. Block intensity comparison code represents the average block intensity difference between blocks of a frame. The extracted feature is further processed using principal component analysis (PCA) to reduce the redundancy while exploiting the correlations between the feature elements. The temporal nature of video is modeled by hidden Markov model (HMM) with BICC as the features. It is found that, BICC outperforms other visual features such as edge, motion and histogram which are commonly used for video classification.     

平行学习——机器学习的一个新型理论框架

本文提出了一种新的机器学习理论框架.该框架结合了现有多种机器学习理论框架的优点,并针对如何使用软件定义的人工系统从大数据提取有效数据,如何结合预测学习和集成学习,以及如何利用默顿定律进行指示学习等目前机器学习领域面临的重要问题进行了特别设计.     

Application of hidden Markov model to acquisition of manipulation skills from haptic rendered virtual environment

This paper explores the feasibility of reconstructing human manipulation skills in complex constrained motion by tracing and learning the manipulation performed by the operator. The peg-in-hole insertion problem is used as a case study, which represents a typical constrained motion force sensitive manufacturing task with the attendant issues of jamming, tight clearance and the need for quick assembly times. In the developed system, position and contact force and torque as well as orientation data generated in the haptic rendered virtual environment combined with a priori knowledge about the task are used to identify and learn the skills in the newly demonstrated task. The recorded training data is classified into contact states, which are identified with hidden Markov model (HMM) as human skills. The HMM parameters are obtained from the training data. By evaluating the controller's performance in each contact state from haptic rendered virtual environment, the robot develops the best trajectories to imitate the human behaviour. In this paper the significance of this research project is highlighted and the developed approach and the progress made so far on this project are reported.     

多注意力集成的图像检索

针对图像检索方法中二阶注意力模块使用全局特征之间的联系所生成的特征存在大量冗余信息,以及集成机制中各分支不能充分训练的问题,提出一种基于多注意力集成的图像检索方法。该方法利用在图像分类任务中表现良好的独立自注意力模块捕捉局部特征之间的联系,生成质量更高的特征以用于图像检索。该方法提出一个多注意力集成框架,在各注意力分支中分别利用独立自注意力模块产生相应的高效图像特征,并通过有效结合产生最终的图像特征。多注意力集成框架利用最终图像特征的排序损失、各注意力分支之间的差异损失及各分支的图像分类损失对模型进行联合训练,使各分支能得到充分训练。在CUB200-2011及CARS196图像检索数据集上的实验表明,所提方法可以有效提高检索精度。     

Human motion capture data retrieval based on semantic thumbnail

We present a method for the efficient retrieval and browsing of immense amounts of realistic 3D human body motion capture data. The proposed method organizes motion capture data based on statistical K-means (SK–means), democratic decision making, unsupervised learning, and visual key frame extraction, thus achieving intuitive retrieval by browsing thumbnails of semantic key frames. We apply three steps for the efficient retrieval of motion capture data. The first is obtaining the basic type clusters by clustering motion capture data using the novel SK-means algorithm, and after which, immediately performing character matching. The second is learning the retrieval information of users during the retrieval process and updating the successful retrieval rate of each data; the search results are then ranked on the basis of successful retrieval rate by democratic decision making to improve accuracy. The last step is generating thumbnails with semantic generalization, which is conducted by using a novel key frame extraction algorithm based on visualized data analysis. The experiment demonstrates that this method can be utilised for the efficient organization and retrieval of enormous motion capture data.     

基于SOM 特性和PCA 索引的三维运动检索

提出了一种基于自组织特征映射(SOM)和PCA 索引的三维运动数据检索方法.首先利用每一个运动序列来进行拓扑特性加强的SOM 的学习,其运动特性被映射到一个主曲面,然后利用主成分分析方法(PCA)提取主曲面的主成分来建立一个基于主成分的索引机制,加快检索速率.SOM 的引入避免了与原始数据的直接接触,后续的工作只是在主曲面的基础上展开,消除了不同骨架长度的位置信息对运动特性的影响.实验结果表明了算法的有效性.