Image Morphing Research And Development

Image morphing is such a technique that could build a series of intermediate frames between two reference images in which content and color may be quite different. As a basic component of IBR and a powerful image manipulation technique, image morphing has been applied to each field extensively, especially in virtual reality, CAD, computer animation, military & police and movie & TV field.View morphing emphasizes that intermediate frames need maintain the shape of reference images relative to image morphing, a basic image manipulation technique. An improved view morphing has been proposed on purpose to match the feature points between two images accurately. First, the single pixel edges of image are extracted by using the algorithm of edge detection, and feature points in the image are found by using method of improved Hough transform. Then, the points are selected accurately as correspondence points across two reference images, and intermediate frames with high quality can be built by the view morphing.Because of some shortcomings about the conventional Hough transform, an improved Hough transform algorithm is presented to improve identifying precision. Line segments are detected by using hierarchical approach, and the point which cumulative number is the most in the local region in parameter space is chosen as the correspondence of a single line in image space. Then, the discontinuous lines can be detected, and the isolated points can be excluded. The endpoints of the line segment can be found too.The in-between images can't maintain the same depth information as the reference image while morphing the complicated images through view morphing algorithm. And then a layered morphing algorithm is given which integrates the reconstruction method of complicated images with image morphing algorithm. First, the reference image is divided into different image layers with depth information, then morphs and reconstructs the in-between images.This paper develops a practical image morphing system, which is based on above various kinds of improved algorithms. This system is proven to be effective by processing different sample images with field morphing, layered morphing, improved view morphing and improved Hough transform.