Object-order versus image-order rendering

For rendering, we need to consider all combinations of objects and pixels. This suggests a nested loop. One way to resolve the visibility is to iterate over the list of all triangles and attempt to render each one to the screen. This is called object-order rendering, and is the main topic of Section 7.2. For each triangle that falls into the field of view of the screen, the pixels are updated only if the corresponding part of the triangle is closer to the eye than any triangles that have been rendered so far. In this case, the outer loop iterates over triangles whereas the inner loop iterates over pixels. The other family of methods is called image-order rendering, and it reverses the order of the loops: Iterate over the image pixels and for each one, determine which triangle should influence its RGB values. To accomplish this, the path of light waves that would enter each pixel is traced out through the virtual environment. This method will be covered first, and many of its components apply to object-order rendering as well.

Steven M LaValle 2020-01-06