Simulation of 3D Virtual Space Model Based on Texture Image & Structure Analysis

In this paper, we employ a 3D reconstruction technique using the multi-view images, i.e., one of the 3D data generation methods. we provide an algorithm for simulating seasonal changes, such as autumn or winter contents, from summer contents in a 3D virtual space model. This algorithm uses two-dimensional (2D) texture analysis and 3D structure analysis in the virtual space model for the simulation. The 2D texture analysis is based on the color and edge components of the texture image. The 3D structure analysis is performed using the normal vector components corresponding to the mesh of the objects. The advantages of the proposed algorithm are as follows. It is efficient in terms of time and cost, as it is not necessary to take pictures every season for producing seasonal 3D contents. This algorithm provides increased efficiency, owing to the small processing computations. It is possible to provide various contents to the user by using only one input source.