Photorealistic scene reconstruction Jankó Zsolt Image and Pattern Analysis Group Computer and Automation Research Institute Hungarian Academy of Sciences A traditional goal of Computer Vision is to compute properties of the three-dimensional world from images. One of the challenging problems of Computer Vision is to reconstruct a three-dimensional model of the scene. Scene reconstruction is a very difficult and complex problem. My research work may be separated essentially into two different parts. At first I worked on the task of reconstructing a 3D model from images. I focused principally on the geometrical part of it. This involves the projection from 3D to 2D, the variants of the camera models, the epipolar geometry, how to use additional information for scene reconstruction (called "stratified reconstruction") and how to get the 3D points from the known 2D point-pairs (triangulation). During the last few months I have been concentrating on the problem of how to present the 3D model as photorealistic as possible; how to combine precise geometric measurements with surface roughness and texture. The future goal is the photorealistic visualization of the objects considering the 3D texture (surface roughness), surface reflectance and shadows.