In this comparative study, we intend to analyse different methodologies to perform 3-Dimensional modeling and printing, by using raw images as input without any supervision by a human. Since the input consists of only raw images, the foundation of the methods is finding symmetry in images. But the images that seem symmetric are not symmetric due to the perspective effect and utterance of other factors. The method uses factors like depth, albedo, point of view, and lighting from the input image to formulate 3D shapes. A 3D template model with feature points is created, and by deforming the 3D template model, a 3D model of the subject is then reconstructed from orthogonal photos. The number and locations of the proper amount of feature points are derived. Procrustes Analysis and Radial Basis Functions (RBFs) are used for the deformation. Images are then mapped onto the mesh following the deformations for realistic visualization. Characterization of the input image shows an asymmetric cause of shading, lighting, and albedo rendering the symmetry of images. The experiments show that using these methods can give exact 3D shapes of objects like human faces, cars, and cats.
Kaushik, Rohit; Vashisht, Chirag; and Kaushik, Eva
"Graphical Image Rendering: Modeling, Animation of Facial or Wild Images,"
International Journal of Computer and Communication Technology: Vol. 9:
1, Article 10.
Available at: https://www.interscience.in/ijcct/vol9/iss1/10