期刊文献+
共找到2篇文章
< 1 >
每页显示 20 50 100
Feature Patch Illumination Spaces and Karcher Compression for Face Recognition via Grassmannians 被引量:1
1
作者 jen-mei chang Chris Peterson Michael Kirby 《Advances in Pure Mathematics》 2012年第4期226-242,共17页
Recent work has established that digital images of a human face, when collected with a fixed pose but under a variety of illumination conditions, possess discriminatory information that can be used in classification. ... Recent work has established that digital images of a human face, when collected with a fixed pose but under a variety of illumination conditions, possess discriminatory information that can be used in classification. In this paper we perform classification on Grassmannians to demonstrate that sufficient discriminatory information persists in feature patch (e.g., nose or eye patch) illumination spaces. We further employ the use of Karcher mean on the Grassmannians to demonstrate that this compressed representation can accelerate computations with relatively minor sacrifice on performance. The combination of these two ideas introduces a novel perspective in performing face recognition. 展开更多
关键词 GRASSMANNIANS Karcher Mean Face Recognition ILLUMINATION SPACES Compressions FEATURE PATCHES Principal ANGLES
下载PDF
Characterizing Placental Surface Shape with a High-Dimensional Shape Descriptor
2
作者 jen-mei chang Amy Mulgrew Carolyn Salafia 《Applied Mathematics》 2012年第9期954-968,共15页
The human placenta nourishes the growing fetus during pregnancy. The newly developing field of placenta analysis seeks to understand relationships between the health of a placenta and the health of the baby. Previous ... The human placenta nourishes the growing fetus during pregnancy. The newly developing field of placenta analysis seeks to understand relationships between the health of a placenta and the health of the baby. Previous studies have shown that the median placental chorionic shape at term is round, and deviation from such prototypical shape is related to a decreased placental functional efficiency. In this study, we propose the use of a nearly-continuous shape descriptor termed signed deviation vector to systematically study the relationship between various maternal and fetal characteristics and the shape of the placental surface. The proposed shape descriptor measures the amount of deviation along with the direction of the deviation a placental shape has away from the shape of normality. Using Linear Discriminant Analysis, we can independently examine how much of the placental shape is affected by maternal, newborn, and placental characteristics. The results allow us to understand how significantly various maternal and fetal conditions affect the overall shape of the placenta growth. Though the current study is largely exploratory, the initial findings indicate significant relationships between shape of the placental surface and newborn’s birth weight as well as their gestational age. 展开更多
关键词 SIGNED Deviation Vector PLACENTA Shape ANALYSIS Linear DISCRIMINANT ANALYSIS Principal Component ANALYSIS
下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部