Research And Application Of Image Deformation Algorithm Based On Regularization Technology

Image deformation has been successfully applied in many different kinds of fields.However,how to get an approach with high efficiency and perfect visual effect remains a challenge task.This thesis mainly studies the image deformation algorithm based on regularization technology and its application: we establish a unified mathematical model for the rigid deformation and non-rigid deformation,and use the regularization term to constraint objective function,transform the deformation problem into vector interpolation problem,introducing image deformation algorithm with vector-field interpolation based on regularization technology;for the problem that the continuous multi-frame image deformation is unnatural,we establish the feature point matching model,followed by feature point fitting,introducing a MEAN-SHIFT automatic deformation method based on regularization technology;in order to solve the problem of recording facial expressions in the film making process,the image deformation application based on regularization technology is introduced,and the simple operation and convenient storage make the expression recording tool extremely practical.The main contributions are as follows:1.Image Deformation with Vector-field Interpolation based on regularization technology.In this thesis,we present a vector-field interpolation method for image deformation,which is based on Moving Regularized Least Squares optimization with Thin-plate Spline(MRLS-TPS).The proposed approach takes user-controlled points as input data,and estimates the spatial transformation for each pixel by the control points.In order to achieve a realistic deformation,we formulate the deformation as a novel closed-form transformation estimation problem by Moving Regularized Least Squares(MRLS).Unlike Moving Least Squares(MLS),we model the mapping function by a non-rigid Thin-plate Spline(TPS)function with a regularization coefficient.Therefore,the deformation can not only satisfy global linear affine transformation,but also adapt to local non-rigid deformation.In terms of the transformation,we derive a closed-form solution and achieve a fast implementation.Furthermore,the approach can show us a wonderful user experience,and give us fast and convenient manipulating.Extensive experiments on 2D images and 3D surfaces demonstrated that the proposed method performs better than other state-of-the-art methods like MLS and the commercial software as Adobe PhotoShop CS 6,especially in the case of flexible object motion.2.Automatic deformation based on regularization technology and MEAN-SHIFT.First,we build a mathematical model,then perform feature point fitting,the MEAN-SHIFT of the regularization technique is used for automatic deformation finally.We use a cascading regression tree for feature point detection,then use MEAN-SHIFT to track and locate the feature points of each frame of the video stream,with the kernel density estimate function approximating the true response map of each frame of image feature points.The regularization technique is used to constrain the deformation function.