Developing sequential methods that could potentially speed up the rendering of special effects in movies.
Computer generated special effects are commonly used in many movies today. One approach to include new objects into a movie or an image is to capture the environment as a HDR image and then use the light from the image to determine the reflections on the new object and to create shadows. This method is therefore known as image-based lightning in the literature. The problem is that this method requires the generation of many thousands or millions of light rays through the scene, which requires a lot of computational resources.
However in a sequence of images, the scene does not change a lot between two frames and there could therefore be beneficial to make use of the information from the previous frame when generating the next frame. Today, the most common approach is to generate each frame independently and this project could potentially be helpful in decreasing the computational cost (and thereby the financial cost) of generating special effects in movies.