Reconstruction of wire frame model of complex images using syntactic pattern recognition
Suman Rani, Deepika Ghai, Sandeep Kumar
Abstract
Reconstruction of 3D images is always a challenging task for the research community and scientist because 3D image reconstruction describes the quantitative characteristics of the objects i.e., shape, size, texture, and volume. The proposed work represents the wireframe model of 3D reconstruction and introduced a new algorithm to represent three-dimensional images in a textual form. In the proposed work firstly identified the wireframe model and then describes the image characteristics. The syntactic based algorithm extracts the image characteristics in the form of knowledge vectors. The proposed work is having a novelty in form of vector of knowledge that provides a significant contribution in security. In addition, the knowledge vector is used to rebuild the image wireframe model and a single pass is enough for the original image to be reconstructed. In the proposed work two standard publicly available database (PASCAL 3D+ and example-based 3D object arrangements) used for evaluation of the proposed work. The accuracy of the proposed method is 93.21%, 94.66% on PASCAL 3D+ and example-based 3D object arrangements datasets. The algorithm takes 1.01 seconds for the reconstruction of the wireframe model. According to the results obtained, the proposed methodology gives better results than the existing state-of-the-art methods.