Axis-Aligned Bounding Box for Image Rendering Sample Test Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image rendering techniques face inefficiencies in sample test efficiency, particularly in motion blur and depth of field rendering, due to high computational costs and low sample test efficiency when handling large numbers of samples and fast motion.
Innovation Solution
The method involves determining the trajectory of a primitive in screen space and constructing an axis-aligned bounding box that overlaps a portion of the screen space region, excluding sample points not overlapped by the bounding box from testing, thereby reducing the number of samples tested for rendering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If time-continuous triangles are used to support motion blur with unique time values per sample, then image quality is improved, but sample test efficiency degrades drastically (can be as low as 1%)
Solution Approach 1:
The patent segments the screen space region into multiple sub-regions using an axis-aligned bounding box that divides the region into tested and non-tested portions. This segmentation allows the system to focus computational resources only on relevant sample points within the bounding box, rather than testing all samples in the entire screen space region, thus improving sample test efficiency while maintaining image quality.
Solution Approach 2:
The patent applies local quality by determining sample test efficiency locally for different regions of the screen space. By constructing an axis-aligned bounding box that overlaps only a portion of the screen space region, the system applies full testing only where the primitive trajectory indicates potential overlap, while excluding samples outside the bounding box. This localized approach maintains high sample test efficiency in relevant areas without unnecessarily testing samples in irrelevant areas.
2Reliability
If all samples in covered pixels are tested by interpolating the triangle to current sample time, then accurate motion blur is achieved, but computational cost increases significantly
Solution Approach 1:
The patent performs preliminary action by constructing an axis-aligned bounding box based on the primitive trajectory before actual sample testing. This bounding box pre-identifies which sample points are worth testing by determining the region where the primitive might overlap during its trajectory. By doing this preliminary filtering, the system avoids the computational cost of testing all samples while ensuring that all potentially relevant samples are included in the test set.
Solution Approach 2:
The patent extracts only the necessary sample points for testing by using the axis-aligned bounding box to identify and exclude samples that cannot possibly be overlapped by the primitive. This extraction process removes unnecessary computational work (samples outside the bounding box) while retaining all samples that might contribute to accurate motion blur rendering (samples within the bounding box).
3Measurement precision
If the number of samples per pixel is increased to improve image quality, then rendering quality improves, but the number of samples requiring testing increases
Solution Approach 1:
The patent segments the set of all samples in a screen space region into two distinct groups: samples within the axis-aligned bounding box that require testing, and samples outside the bounding box that can be excluded. This segmentation becomes increasingly valuable as the number of samples per pixel increases, because the bounding box typically covers only a fraction of all possible samples. The segmentation approach maintains rendering quality by testing all relevant samples while dramatically reducing the total number of tests required.
4Speed
If fast motion is rendered with high sample test efficiency, then rendering speed is improved, but maintaining accuracy becomes more difficult
Solution Approach 1:
The patent introduces another dimension (the time dimension or non-screen dimension) by constructing the axis-aligned bounding box to span the temporal trajectory of the primitive. Instead of only considering spatial overlap in the screen space, the system incorporates temporal information by extending the bounding box across the primitive's path through multiple time points. This dimensional extension allows the system to accurately capture fast motion trajectories while maintaining sample test efficiency by excluding samples that are outside the temporal-spatial envelope defined by the bounding box.
Data Source
AI summary
A bounding box-based method for reducing the number of samples tested for rendering a screen space region of an image includes determining a trajectory of a primitive in screen space for an image which is to be rendered and constructing an axis-aligned bounding box for the screen space region. The axis-aligned bounding box includes a bound in a non-screen dimension that is defined as a function of the screen space trajectory of the primitive, and overlaps a portion of the screen space region. One or more sample points which are located within the screen space region, and which are not overlapped by the axis-aligned bounding box are excluded from testing.