Síntesis de imágenes a partir de imágenes reales de una escena mediante un algoritmo parcialmente autónomo

This paper presents an image synthesis algorithm for rendering views of a real scene taken with a virtual camera which is located between two real cameras. Specifically, this paper presents how, under certain conditions, a pair of views is enough to obtain a full set of possible views by following the line that joins their optical centers (baseline) without reconstructing explicit 3D models. Test results show the viability of the proposed algorithm for simple scenes (an object with a simple geometry, which is opaque, and does not present any occlusions).