1) projected branch
投影分支
1.
In this paper, A new concept of projected branch is introduced, and a new algorithm FTPB (frequent subtrees mining based on projected branch) is proposed.
在频繁子树挖掘中引入投影分支的概念,并提出基于投影分支的快速频繁子树挖掘算法——FTPB。
2) projecting integral
投影积分
1.
The image of the measuring mark of plate captured by CCD,recognized and analyzed the registered mark by the statistic character of hue in HSI space,and segmented the color image by the 2 paramaters of hue and intensity,measuring data about the registering quality can be obtained by hue projecting integral method after color image segmentation.
应用CCD采集印版测量标识的图像,在HSI色彩空间中,通过色调统计特征进行套准识别,并通过色调和亮度对彩色图像进行阈值分割,对分割后的图像采用色调投影积分方法快速获取测量标识的坐标参数,实现彩色套印误差的检测。
3) integral projection
积分投影
1.
Eye location algorithm based on differential and integral projection;
一种微分与积分投影相结合的眼睛定位方法
2.
Design and implementation for integral projection Vector Quantizer.;
一种基于积分投影的矢量量化器的设计与实现
3.
Fast eye and mouth location algorithm based on integral projection and color matching
一种基于积分投影与色度匹配相结合的快速人眼嘴定位算法
4) Integration projection
积分投影
1.
According to the characteristics of vehicle windows,we employed Hough transform and Integration projection to cut the window regions.
针对图像中车窗边缘的图像特点,提出了一种基于相位编组法进行图像分块,在图像块内进行快速Hough变换的直线检测,并结合积分投影方法对车窗进行定位与提取。
2.
A possible position and the size of a face were predicted through integration projection.
该算法利用积分投影法预测人脸可能的范围和位置,并结合椭圆的紧密度概念调整滤波器参数,构造最优滤波器,把肤色分割后的连通区域中非人脸的类肤色区域过滤掉,有效地减少类肤色区域带来的干扰,再结合模板匹配方法对人脸进行定位。
5) differential projection
微分投影
1.
The presented algorithm combines the traditional integral projection method and the differential projection method which is proposed in the paper.
文章介绍了一种精确定位眼睛的方法,该算法将眼区灰度总体分布特点与眼部灰度变化特点相结合;将传统的积分投影方法与本文提出的微分投影相结合;实验结果表明,该算法对光照变化不敏感,定位准确率高。
2.
The presented algorithm combines the traditional integral projection method and the differential projection method,which is proposed in the paper.
介绍了一种精确定位眼睛的方法,该算法先利用改进的人脸图像的垂直灰度积分投影确定脸的左右边界,再根据人脸图像的水平灰度投影曲线来确定眼睛的大致高度;还将传统的积分投影方法与本文提出的微分投影相结合,实验结果表明该算法消除了背景、头发及服装等干扰,定位准确率高。
6) projective distribution
投影分布
1.
One sufficient and necessary condition is obtained for normal projective distribution through generalization of the conclusion in the literature[1], which provides a fact of abnormal population density in n-dimensional Euclid space and normal projective density in any proper subspace by means of projective search.
文献[1]通过投影寻踪方法得到了一个在n维空间的总体密度为非正态,而在任意维真子空间上的投影分布为正态的例子。
补充资料：滤波反投影或卷积反投影
滤波反投影或卷积反投影
影像学术语。当代影像学设备进行影像重建的数学方法。在直接用扫描后所获得的投影轨迹剖面图反投影重建出的CT图像中，无法避免角度卷入条纹伪影(angular aliasing streaks)造成的模糊和失真。这种现象与被扫描层面的空间频率中高频信息的损失有关。使用一种精密的数学方法去除这种模糊。称为“展现”(unfolding)或去卷积(deconvolution)，即在反投影前使用一种数学的“滤器”或卷积函数对原始数据进行修正，然后再进行反投影。两步数学处理过程合称为滤波(修正后)反投影或卷积(后)反投影。这种方法的优点是处理过程简单，速度快，所得图像逼真、清晰。
说明：补充资料仅用于学习参考，请勿用于其它任何用途。
参考词条