Non-photorealistic rendering in computer graphics

• Main Author: Chan, Hui San.
• Other Authors: Henry Johan
• 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/10356/46344
• Institution: Nanyang Technological University

Non-photorealistic rendering is one of the popular digital image processing techniques in computer graphics that is able to generate expressive images in forms of painting, sketches, animation, illustrations and etc. The objective of this project is to focus on 2D non-photorealistic rendering (NPR) to automatically generate an artistic image from an input 2D colour image. This report consists of four image processing techniques used in generation of artistic image namely: Greyscale Filter, Sobel Operator, Sharpening using Convolution and Gaussian Blur using Convolution. Sobel operator is one of the image processing techniques discussed during the project to detect the edges and outline of an original image and it is the process of emphasizing the edges. Then, Sharpening with convolution is also examined and computed on the input image to sharpen the image so as to emphasize and highlight the fine details of the image. Lastly, Gaussian Blur with Convolution is discussed and applied on the image to achieve blurring effects on the original image. Artistic images are then generated by having different combinations of the implemented filters. The various output effects after generating artistic images from the above mentioned image processing techniques are also discussed and explained in this project report. At the end of the project, input images are successfully transformed to artistic images with the use of corresponding image processing methods. This report is divided into chapters and each chapter describes in detail the development process which was carried out during this project period.