Font Size: a A A

Human Animation And Face Exaggeration Using Image Deformation

Posted on:2013-02-20Degree:MasterType:Thesis
Country:ChinaCandidate:W H ChenFull Text:PDF
GTID:2218330371477785Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
A.s the development of the information society, the widely use of multimedia computer technology are paid more and more attention by people. Image processing which is one of the most significant technologies of multimedia becomes a hot research field in these years. Meanwhile, image processing relates to many kinds of technology and has many sub-directions. Image deformation is an important subfield. Our thesis is to research and improve the classical image deformation algorithm for a better realistic and real-time result. And combining with other technologies in image processing, we apply the algorithm on our lives and even industry. The details are listed below:1. Achieving the deformation of2D human body image and getting an animation result. Our thesis improves the image deformation algorithm in two ways. One is to use a variable weight function, instead of a constant weight, which can avoid the distortion ion the course of deformation. The other is to change the influence area of the control points in the deformation and increase the speed of calculation that keep the deformation real-time. This improvement would give some advice for today's cartoon industry which still use the hand-painted key frames.2. Achieving the face exaggeration. Our thesis combines the image deformation algorithms with face feature detection and maximally stable extremal regions. Using these algorithms, we can detect and deform the feature in human face and give them an exaggeration effect or a better look. Today, more and more people devote to photographic industry, and what they used for their photo processing is Photo Shop. But we offer them a more automatic and easier way to process their face-image.。...
Keywords/Search Tags:MLS, Image Deformation, ASM MSER
PDF Full Text Request
Related items