Abstract
We propose a method to compose the image of an arbitrary viewpoint from the eight cameras array. The synthesis is based on the reconstruction filter derived from the light field rendering. The image composition can be implemented in real time using an FPGA.