Modeling and rendering architecture from photographs: a hybrid geometry- and image-based approach
SIGGRAPH '96 Proceedings of the 23rd annual conference on Computer graphics and interactive techniques
Theory and Practice of Projective Rectification
International Journal of Computer Vision
Multiple View Geometry in Computer Vision
Multiple View Geometry in Computer Vision
Interactive image completion with perspective correction
The Visual Computer: International Journal of Computer Graphics
| Hi-index | 0.00 |
Construction of 3D models from 2D images is yet challenging as achieving photo realism with accurate representation of objects is complex. The users, while navigating the 3D scene may not feel the reality in the 3D models when the objects are not represented as in the real world. The reason for most of the misrepresentation is improper computation of 3D coordinates and construction of wireframes which are the building blocks of the 3D scenes. In this paper, a method is proposed to represent the 3D coordinates more precisely which will result in more photo realistic models. The approach is carried out in two folds. First each surface of planar objects is rectified for perspective distortion using plane homography. Then based on the width and height of each of the surfaces, a wireframe is drawn. The orientations of surfaces are assumed to be orthogonal to each other. The 3D coordinates of each surface is computed based on the orientation and dimension (width and height) of each surfaces. Keeping the very first coordinate of the scene as reference, other coordinates of all surfaces are computed very precisely. This approach is compared with our previous method based on depth cueing which is constructs the wireframes approximately. The experimentation on objects of planar surfaces like buildings, floors and walls had proven that this method is robust for image based modeling.