其次，在曲面片内部，曲面拟合点和边界方向导数可以从参考曲线中通过构建虚拟等参线的方法得到。
Next, in the interior of the patches, surface fitting points and cross boundary derivatives are estimated from the reference curves by constructing virtual iso-parametric curves.
我们已经提到了一点有关算法可能会与训练集过拟合（over-fit）的细节。
We've already talked a bit about the fact that algorithms may over-fit the training set.
回归软件程序将样本点用图表示，并且给出最拟合样本点的直线的价值函数公式。
Regression software programs graph the sample points and give the (cost-function) formula of the straight line that best fits the points.
前向网络利用反向传播算法训练多层网络，使训练后的网络较好地拟合样本空间中各点的函数值。
Feedforward networks use back propagation algorithm to train a multi-layer network. After training, the multi-layer network can fit the function in the data space very well.
然后基于规则筛选出关键特征点集，通过两步最小二乘法拟合牙弓线。
Thirdly, the feature point set is filtered by the rule and two-step curve fitting method is used to obtain the dental arch form.
然后，着重讨论了用于轴承边界的圆度、尺寸和圆心度检测的两种计算方法—点—圆变换法和圆形曲线拟合法。
Then, two methods are discussed which are point-circle transformation and circular curve fitting for inspecting the bearing's circularity, dimension and concentricity.
提出一种新的轮廓特征点提取方法，它利用最小二乘法拟合轮廓曲线，结合斜率、曲率信息提取特征点。
A new feature-points extraction algorithm is proposed, which USES least square method fitting contour curve and combines slope, curvature information to extract feature-points.
该测量方法具有测量采集点精度高、齿形轮廓曲线拟合误差小、测量过程与误差处理过程人工干预少、测量精度高等特点。
The method has higher precision of getting measuring points, less fitting error of tooth form, less artificial interpose in the measuring process and higher measuring precision.
利用相关系数在真实匹配点附近的统计特性，通过对相关系数进行曲线拟合，根据其一阶导数确定最大相关系数所在的位置，快速得到亚像素匹配点。
By using the statistic properties of the correlation coefficient, a curve correlation coefficient is calculated, and then the matching point is obtained through evaluating it's derivative function.
通过对流体力学中沿程阻力系数和泵与风机工作点等问题的求解，介绍了EXCEL单变量求解器和曲线拟合的方法。
Through calculating friction factor and operating point of pump or fun of two examples on fluid mechanics, methods using solver and curve-fitting in EXCEL were introduced.
研究了用圆弧曲率估算截面数据离散曲率的方法和特征点提取方法，提出了基于曲线拟合的截面特征识别方法。
Other preprocessing methods such as discrete curvature estimation and feature point extraction are also studied, and a section-feature recognition algorithm, based on curve fitting, is proposed.
本文研究了样条曲线拟合中常遇到的两种畸变点的处理方法。
The approaches to the treatment of two kinds of distortion points often encountered in spline curve fitting are investigated in this paper.
但二次多项式只能拟合x坐标与Y坐标之间成函数关系的点序列，无法拟合闭合曲线。
But the quadratic curve can be used only in condition when each X-coordinate of the points sequence mapping to only one Y-coordinate, so it can't be used in fitting closing curve.
通过对离散点进行短线段拟合的基础上，再利用曲线拟合的最小二乘法对圆和直线进行二次拟合。
On the basis of short straight-line fitting, this dissertation USES the least-square method of curve matching to discern circle and straight line.
最后应用神经网络，反复对边缘特征点进行拟合训练，得到平滑的边缘拟合曲线。
Finally, a smoothing fitting curves will be obtained by training of the edge feature using of neural networks.
本文在这两种方法的基础上提出用样条函数拟合纸上定线中的离散点，并给出公式推导和示例。
On the basis of these, this paper uses mode function to imitate the disperse dots, and give formulas and examples.
数据拟合问题在日常生活中经常遇到，点的拟合可以通过插值等方法来实现。
Data fitting is frequently used in our life. The points fitting process can be achieved through interpolation methods.
采用在不同时间点增加阻力系数修正系数的方法拟合速度-时间曲线和位置-时间曲线。
Different drag correction coefficients at different time points are introduced, and then both speed curve and ballistic curve fitting are finished.
受经典物理学中极性场模型及扩展点集思想的启发，提出一种称为“场拟合”的隐曲面重建方法。
Motivated by the concept of classic physical polar field model and off-set points strategy, we present a new approach, called field fitting.
其中曲线拟合是一个在给定误差下进行曲线逼近的过程，当曲线拟合完成之后，整个曲面的控制点的数目也就确定了。
Curve fitting is a curve approaching process under definite error and the number of control points throughout surface is ascertained after curve fitting completed.
结果表明采用曲线拟合法能使测距绝对误差在3个采样点以内，与传统的反射波起始点确定方法相比有较大优势。
The absolute error is no more than 3 sampling points, so it has more advantages compared with the traditional methods to judge the start point of reflect wave.
目前灰色系统理论在建立GM（1，1）模型时通常采用假定拟合曲线通过建模数据第一点来确定积分常数，从而得到预测公式的方法。
When grey systematic theory at present set up GM (1 , 1 ), it assumes that fit curve passes the first point of modeling data to confirm the integral constant, thus obtained to forecast formula.
以曲率极大点作为初始拟合点。
The vertices with maximal curvature are as the initial fitting vertices.
将中心点的空间距离作为一个拟合参数引入预测过程，提高了预测精度和消噪能力。
The prediction precision and the ability of noise elimination are improved by introducing space distance of central point as fit parameter into predicting process.
超精密数控车床加工精度的提高不仅依赖于数控装置的运算速度和曲线的拟合精度，同时还依赖于伺服系统的跟踪精度和点位控制精度。
The manufacture precision of super exactitude NC lathe depends on not only the operation rate of NC deviceand joined precision, but also the error of track and point location about servo system.
因此，这里采用三次样条函数拟合曲线构造模具型腔廓线，其型值点由已有模具型腔测得。
So it adopt cubic spline function to structure die cavity outline in this article, the model points may be surveyed on die cavity surface.
曲面内腔类零件的反求有别于轮廓外形的反求，主要体现在数据采集的方法，点云的处理，特征点的提取以及曲线的拟合。
The reverse of lumen part is different from outline, such as collect data method, treat with point, pick up feature point and curve fitting.
根据各特征点对轮廓的贡献，采用弦长加权拟合法对轮廓点进行处理，从而得到轮廓的中心位置。
The points were set different weight for their different effect on the center of the contour in fitting process.
根据各特征点对轮廓的贡献，采用弦长加权拟合法对轮廓点进行处理，从而得到轮廓的中心位置。
The points were set different weight for their different effect on the center of the contour in fitting process.
应用推荐