Adaptive 2D To 3D Image Conversion Using A Hybrid Graph Cuts And Random Walks Approach

This thesis proposes an adaptive method for 2D to 3D conversion of images using a user-aided process based on Graph Cuts and Random Walks. Given user-defined labelling that correspond to a rough estimate of depth, the system produces a depth map which, combined with a 2D image can be used to synthesize a stereoscopic image pair. The work presented here is an extension of work done previously combining the popular Graph Cuts and Random Walks image segmentation algorithms. Specifically, the previous approach has been made adaptive by removing empirically determined constants; as well the quality of the results has been improved. This is achieved by feeding information from the Graph Cuts result into the Random Walks process in two different ways, and using edge and spatial information to adapt various weights. This thesis also presents a practical application which allows for a user to go through the entire process of 2D to 3D conversion using the method proposed in this work. The application is written using MATLAB, and allows a user to generate and edit depth maps intuitively and also allows a user to synthesize additional views of the image for display on 3D capable devices.