Flash better : image vectorization for animation

Abstract

The primary concern of this project is to analyze the various image vectorization techniques which have already been proposed, to determine which would be the most efficient method to be utilized in the vectorization of line drawings and finally attempt to recreate the results as presented in the research papers. We analyzed methods that were used in the vectorization of construction drawings, handwriting recognition and also methods like Halftone Dots, Triangulation, Optimized Gradient Meshes, Beizer Curve Representation and B-spline Curve Approximation. In the comparison of the techniques, we looked at a few factors. They were the space requirements of the technique, speed of the technique, processor requirements, accuracy and the final results as presented in the research papers. Upon completion of the study, an attempt to realize the chosen method was made. However this proved to be unsuccessful as the objective was deemed too complex and also due to time constraints. In future work, it is advised to further improve the factors taken into account and to look at other methods that may have been further researched after the completion of this project.