مشروع البحث: دراسةحول استخدام تحليل المكونات الاساسية في تحليل الشكل الاحصائي مع بعض التطبيقات
تحميل...
المساهمين
الممولين
رقم التعريف
24321
الباحث
أريج عبدالناصر سليم
الوصف
Abstract
This thesis aims to investigate the modeling of shape variability using Principal Component Analysis (PCA). PCA IS well-known AND WIDELLY USED TECHNIUE IN CLASSICAL MULTIVARIATE. Statistical Shape is another type of the multivariate data therefore, the adapting and implementing on a large number of matrices with a large set of PCA is possible. The shape data to be meaning and sensible is characterized by large number of involved variables and less number of objects i.e. small sample. Also each object or sample member is represented as vector of points. Such points usually of 2D and possibly 3D. The resulted data matrix will be huge and therefore PCA as data reeducation in multivariate analysis still works in this case. On other word, in summarizing (reducing) a large set of variables into a smaller set of unrelated (orthogonal) variables than the original variables. implementing PCA in shape analysis is considered a modern method in contrast to the classical method of PCA. We notice that its steps are not tiring although are more compared its counterpart in classical one. We also notice a difference in interpretations. We note that there is great interest in statistical shape analysis from the point of view of image analysis. In statistical shape analysis the employing PCA requires the use of Procrustes analysis. in this thesis such employing was illustrated and discussed in depth with two different real data sets treated as shape data. The first data set is taken from image recognition and it represents "3" digit. The second example of anatomy is related to or useful for anatomical analysis and it represents the shape of the hand.
الكلمات الدالة
تتميز بيانات الشكل بعدد كبير من المتغيرات المعنية