View Morphing Technology Research

Morphing techniques combine 2D interpolations of shape and color to create dramatic special effects. Despite being computed by 2D image transformations, effective morphs can suggest a natural transformation between objects in the 3D world.In this paper, we present an extension called view morphing that allows current image morphing methods to easily synthesize changes in viewpoint and other 3D effects. When morphing between different views of an object or scene, the technique produces new views of the same scene, ensuring a realistic image transition. View morphing works by prewarping two images, computing a morph (image warp and cross-dissolve) between the prewarped images, and then postwarping each in-between image produced by the morph. Any of image morphing techniques, may be used to compute the intermediate image.To accomplish the image interpolation automatically, this paper proposes a robust approach to match features by exploiting the available constraints. Firstly, the feature points are extracted by Harris algorithm. The matching idea underlying our approach is to use some classical techniques, including correlation and relaxation methods, the Least Median of Squares and 2D projective warping, to find an initial set of matches. Then the epipolar geometry constraint and the 2D projective constraint can be accurately estimated using a well adapted criterion. More matches are eventually found by using the recovered constraints. Using these matched points the image area is divided into a few of triangle areas. The change of a triangle can be designed intentionally to control image deformation inside it. Then a corresponding relation of color is established between a new pixel after triangle change and original image pixels. The quality of the resulting image is demonstrated with interpolations of real imagery.