Image rendering from the depth layer of a scene

• Main Author: Dang, Kang
• Other Authors: Chan Kap Luk
• Format: Final Year Project
• Language: English
• Published: 2011
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/45225
• Institution: Nanyang Technological University

In this exciting and challenging project, we have explored new possibilities of applying artistic rules into a landscape photograph by considering a depth-based partition as one of the input. Compared with other depth rendering techniques, our project not only can enhance the perception of depth, but also take artistic elements into consideration such as emphasizing the overall structuring and layout of images. Our project takes a template-based approach. User selects a photograph, gives the depth-based partition and selects a painting whose style would be used as a template. Rendering algorithms will generate an output image based these inputs and artistic rule. It is easy to identify three key issues: (1) partitioning of image according to its depth (2) extraction of rendering rules from the artistic principles (3) development of rendering algorithms for transferring the style of paintings to photos. These three issues have been thoroughly discussed in this report. In addition, a Graphics User Interface has been programmed to guide the user through the entire procedure and implementation is also mentioned in the report.