Removing Camera Shake Effect from a Photograph

Abstract

This project develops a method to remove the blur effect due to camera shake in a photograph. All the tasks that carried out during the project can be characterized into two areas. First is the development of a software framework for the algorithm development and testing. It includes tools to visualize pixel patterns better and support Implement, experiment, and evaluate various algorithms. The second Item is the development of the main methodology, which is the invention of this project, to remove the camera shake effect from a photograph. The methodology was an evolutionary development. It involved a great deal of experiments, analysis of results and feedback. The mechanism includes Pre Processing, Blurred Edge Detection, Edge Correction and Post Processing. All these functions have got evolved throughout the project, in order to produce better results. In order to support all the experiments done for the main algorithm development, a rich set of tools and components were needed in the supporting framework. Standard graphs such as color histogram, gradient histogram, together with some custom graphs, are implemented. Standard image processing algorithms like Blur kernels, Median Filtering and Frequency Domain Analysis was implemented. Lot more other tools were developed to visualize the pixel patterns better.