Click slide for next, or goto previous, first, last slides or back to thumbnail layout.
Click slide for next, or goto previous, or back to thumbnail layout.
Click slide for next, or goto previous, first, last slides or back to thumbnail layout.
Click slide for next, or goto previous, or back to thumbnail layout.
This slide gives an overview how view-based rendering works.
First, three views are chosen based on the viewing direction, using the same mechanism as presented before. The textured mesh of each of the three views is rendered from the current viewpoint.
The final image is composited from these three renderings pixel by pixel. The color and distance of the three input pixels are first read in. If some of the inputs are clearly occluded by others, the occluded ones are removed. The remaining ones are blended together using the three weights, and the resulting color is assigned to the pixel. The same process is repeated for each pixel.