Creating special image effect by image decomposition and reconstruction

• Main Author: Li, Mengguo
• Other Authors: Jiang Xudong
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Computer hardware, software and systems
• Online Access: http://hdl.handle.net/10356/64394
• Institution: Nanyang Technological University

In recent years , face image processing has become a popular research field. There are many successful app lications based on this image analysis techni que. Therefore , face recognition and creating speci al effects is widely used base on the characteri stics. Image processing consists of digi tal imag e processing and analog image processing , Digi tal image processing performs images on two-dimensional (2-D) by computer algorithms. And analog image processing performs images by analog signals. Most of lime, we will use digital image processing instead of analog image processing because it can allow much wid er range of algorithms applications to the input image and also can avoi d pro blems like incre asing noise and sign al distortion during processing. In this project, we will use Principal Component Analysis (PCA) to conduct digital image decomposition and reconstruction for human face images. Princi pal Component Anal ysis (PCA) was first invented by Karl Pearson , as an analogue of the principal axes theorem in mecha nics, and was later independ ently developed (and named ) by Harold Hotelling in the 193Os.(I] II is widely used for statistical algorithm to do data representation , feature extraction, and compression in the field of face image processing. And it can increase the efficiency of data processing procedure, which is very helpful in real-life algorithm implementation. The basic concept of PCA is to reduce the large dimensionali ty of observations of strong correl ated variables to the smaller intrinsic dimensionality of a set of values of linea rly uncorrelated variables called principal compo nents. Because of this property , PCA is the best method to extracts the eigenvectors from a covariance matri x constructed from an image database.