基于深度学习的人体姿态估计方法综述

基于深度学习的人体姿态估计方法旨在通过构建合适的神经网络，直接从二维的图像特征中回归出人体姿态信息。主要按照2D人体姿态估计到3D人体姿态估计的顺序，并从单人检测与多人检测、稀疏的关节点检测与密集的模型构建等方面，对近年来基于深度学习的人体姿态估计方法进行系统介绍，从而初步了解如何通过深度学习的方法得到人体姿态的各个要素，包括肢体部件的相对朝向和比例尺度、骨骼关节点的位置坐标和连接关系，甚至更为复杂的人体蒙皮模型信息。最后,对当前研究面临的挑战以及未来的热点动向进行概述，清晰地呈现出该领域的发展脉络。     

A method of identifying latent human errors in work systems

Work system improvements are implemented in various manufacturing processes to prevent problems caused by human errors. However, they are almost always applied to problems which have already occurred. This paper examines a method of identifying latent human errors existing within the work systems beforehand. A procedure for applying failure mode and effect analysis to this identification problem was defined based on over 1000 empirical errors: a work system decomposition criterion and fundamental error modes for listing latent human errors, and then applied to three practical manufacturing processes in order to evaluate its effectiveness.     

等偏移距自动初至静校正

茅金根 ,梁秀文 ,杨午阳 ,冯有奎 .等偏移距自动初至静校正 .石油地球物理勘探 ,2 0 0 2 ,37(5 ) :4 6 9～ 4 72　　本文提出一种自动计算静校正量的新方法。它首先在等偏移距道集内对相邻道的初至波采用高精度相关构建一个关于炮、检时差的超大矩阵方程 ,然后采用优化统计迭代算法求解炮点和检波点的静校正量 ,并用概率分布函数确定精确的炮、检静校正量 ,最后用误差校正函数消除炮、检静校正误差。实际资料处理结果表明 ,本方法自动、省时、精度高。     

Random errors caused by temperature in magnitude of principal strains evaluated ith 3-element strain gauge rosettes

The effects of the uncertainties associated ith the apparent strain and gauge factor data, given by the manufacturer, in the evaluation of principal strains are examined. Principal strains are obtained by a reduction of the strains measured ith electrical resistance strain gauge rectangular or delta rosettes. The theoretical analysis points out the relevance of the effects caused.     

非透视3点问题的解析解

由于在机器人导航等领域新的图像装置不符合透视规则,无法运用传统的透视位置估计方法,采用吴氏零点分解方法,给出非透视3点问题的解析解.利用新图像装置的几何条件,建立了非透视3点方程系统.     

Stabilization techniques for spectral methods

We introduce a stabilized treatment of spectral methods. The condition number of the spectral systems is highly improved. Elliptic and biharmonic problems are considered. Suitable interpolants in the case of inhomogeneous Dirichlet boundary conditions are presented. For a direct solver the improvements with respect to rounding error propagation are numerically demonstrated.     

论滚齿加工在机测量技术

本文论述了滚齿在机测量中一些重要问题．应用偏心误差理论，揭示了以齿轮的齿顶、齿根和内孔为测量基准的测量齿距误差的实质，并对这三种基准进行了对比分析．指出了不同测量基准所测结果不同，若基准选择不当，有时产生“误收”或“误废”．论文针对当前滚齿在机测量装置存在问题．提出了一种滚齿在机１８０°转位双啮综合测量法，测量原理正确，结构简单，易于推广．     

A Multi-View Face Recognition System

In many automatic face recognition systems,posture constraining is a key factor preventing them from application.In this paper a series of strategies will be described to achieve a system which enables face recognition under varying pose.These approaches include the multi-view face modeling,the threschold image based face feature detection,the affine transformation based face posture normalization and the template matching based face identification.Combining all of these strategies,a face recognition system with the pose invariance is designed successfully,Using a 75MHZ Pentium PC and with a database of 75 individuals,15 images for each person,and 225 test images with various postures,a very good recognition rate of 96.89% is obtained.     

Robot Pose Estimation in Unknown Environments by Matching 2D Range Scans

A mobile robot exploring an unknown environment has no absolute frame of reference for its position, other than features it detects through its sensors. Using distinguishable landmarks is one possible approach, but it requires solving the object recognition problem. In particular, when the robot uses two-dimensional laser range scans for localization, it is difficult to accurately detect and localize landmarks in the environment (such as corners and occlusions) from the range scans.In this paper, we develop two new iterative algorithms to register a range scan to a previous scan so as to compute relative robot positions in an unknown environment, that avoid the above problems. The first algorithm is based on matching data points with tangent directions in two scans and minimizing a distance function in order to solve the displacement between the scans. The second algorithm establishes correspondences between points in the two scans and then solves the point-to-point least-squares problem to compute the relative pose of the two scans. Our methods work in curved environments and can handle partial occlusions by rejecting outliers.