Anti-aliasing method based on rotated spatial filtering

• Ano de defesa: 2021
• Autor(a) principal: Monteiro Júnior, Lázaro Ribeiro
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• dARK ID: ark:/74562/001300000322m
• Idioma: eng
• Instituição de defesa: Centro de Ciências Exatas e Naturais - CCEN; Brasil; UFERSA; Universidade Federal Rural do Semi-Árido; Programa de Pós-Graduação em Ciência da Computação
• Palavras-chave em Português: CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO; computer graphics; antialiasing; spatial filtering; linear regression; computação gráfica; filtragem espacial; regressão linear
• Link de acesso: http://lattes.cnpq.br/5111916887378777; http://lattes.cnpq.br/8728455456208244; https://repositorio.ufersa.edu.br/handle/prefix/12035

Due to the evolution of graphics processors over the last decades has become possible to produce highquality and realistic threedimensional scenes. However, aliasing is produced during the sampling performed in the rasterization process. It causes a serrated effect on the edges of the objects presented in the scene, highlighting the unreal aspect of the image and displeasing the viewer. The present thesis aims to develop an antialiasing treatment based on rotated spatial filtering. It performs edges recognition applying spatial filtering along with simple linear regression technique. Then, a smoothing spatial filter is rotated to match the direction of the inspected border and applied in the affected regions. The testing was performed on an OpenGL application, processing the rendered image from the framebuffer. Later, the application was modified to process images from Blender, a 3D modeling software that allows more complex graphics scenes. The results show the proposed method’s effectiveness in smoothing aliasing with good quality and preserving the scene details. Hence, the problem was handled effectively with a postfiltering approach and without oversampling. The algorithm’s execution time is O(n), and the memory used is O(n)