Image processing (image enlargement)

Abstract

Image enlargement is an important aspect in digital image processing. To resize an image, interpolation was used. However, the classic methods have the problems of blurred edges or artifacts around edges. Also the computation time is a problem in some of these methods. In this project, a new interpolation algorithm was developed base on pixel level datadependent triangulation. Using this new algorithm, high resolution image can be created from a lower resolution image. The computation cost and the memory requirement are low. The quality of enlarged image using the new method was accessed visually and by MSE measure, which shows the qualities of the interpolated images using the new algorithm are better than the traditional methods. Edges are sharper while the other areas are smoother.