Anisotropic filtering with using texture maps of weight coefficients

• Дата: 2017
• Автори: Romaniuk, O. N., Dudnyk, O. O.
• Формат: Стаття
• Мова: Ukrainian
• Опубліковано: Інститут проблем реєстрації інформації НАН України 2017
• Теми: anisotropic filtering; Gaussian pixel model; texture; texture mapping
• Онлайн доступ: http://drsp.ipri.kiev.ua/article/view/126539
• Назва журналу: Data Recording, Storage & Processing

The application of anisotropic texture filtering for the rendering of images with a high degree of realism is analyzed. It is justified the need of using weighting factors that take into account the physical characteristics of the pixels output from averaging colors texels.It was found that among all models of pixel in the most adequate is the Gaussian model according to which, a pixel in screen space has the shape of a body of rotation of a Gaussian curve with a base diameter equal to 1 and the volume is equal to 1. The formulas for use Gauss model when performing anisotropic filtering are provided.It is shown that the high computational complexity makes it difficult to use weighting functions based on Gauss model of a pixel in complex graphics scenes with high requirements for computing performance.To reduce computational complexity calculation of weight functions in the process of anisotropic texture filtering, a method of determining the weights by using special texture maps has been proposed. It is shown that the application of texture maps gives the possibility to significantly improve performance of anisotropic texture filtering through the use of pre-calculated weighting factors.