Automatic shader filtering method, device and system based on function approximation of definition domain and value domain

Abstract

The invention discloses an automatic shader filtering method, device and system based on function approximation of a definition domain and a value domain. The method comprises steps of a shader beinganalyzed into a plurality of computing nodes organized by an abstract syntax tree; dividing all computing nodes on the abstract syntax tree into a plurality of target sub-trees according to a data flow direction, wherein each target sub-tree comprises a plurality of adjacent computing nodes; for each target sub-tree, performing fitting function approximation fitting on a drawing calculation function corresponding to the target sub-tree, and calculating an input median and an input variance according to the determined fitting function, and outputting the median and an output variance in a definition domain of the drawing calculation function, wherein the input median and the input variance are the output median and the output variance of a previous target sub-tree connected with the currenttarget sub-tree in the data flow direction; and taking the output median of the final target sub-tree as an automatic filtering result of the shader to improve the universality, efficiency and quality of shader filtering, thereby improving the quality and application range of an anti-aliasing technology.

Description

technical field [0001] The invention relates to the technical fields of graphics drawing, signal filtering and automatic testing, in particular to a shader automatic filtering method, device and system based on a function approximation of a definition domain and a value domain. Background technique [0002] During the drawing process, the screen pixel color can be understood as a discrete sampling of the continuous signal on the surface of the drawn object. According to the signal sampling theorem, only when the signal sampling frequency is greater than twice the highest frequency in the signal, can the sampled digital signal fully retain the information in the original signal. +Due to the constraints of real-time rendering on efficiency, the sampling frequency is not enough, resulting in common aliasing. In graphics, the technique used to reduce or eliminate this effect is called anti-aliasing. When displaying raster graphics, it can be understood as sampling continuous s...

AI Technical Summary

Problems solved by technology

However, this method may be severely distorted during the convolution process, because the iterative process takes into account the domain and value range effects, and the results may appear outside the value range during the calculation of the median and variance

[0005] Therefore, filtering for arbitrary shaders remains a challenging problem

Images