Interactive modeling of lofted shapes from a single image
- PDF / 11,681,190 Bytes
- 11 Pages / 612 x 808 pts Page_size
- 60 Downloads / 220 Views
Interactive modeling of lofted shapes from a single image Congyue Deng1 (
), Jiahui Huang1 , and Yong-Liang Yang2
c The Author(s) 2019.
complicated geometry. Moreover, to achieve the level of detail required for realistic objects, modeling often proceeds via iterative refinements requiring significant effort. The above factors limit the applicability of traditional modeling techniques, and motivate research into more efficient tools which can make models from a single image. This can not only bridge the gap between having a vision of a 3D object and modeling its geometry, but also benefit novices by providing intuitive control over the results. The cognition of geometric shapes from a single image is an easy task for humans who have profound knowledge of sensing the 3D world. Nonetheless, 3D shape reconstruction from a single image is an illposed problem in general, due to the lack of depth cues and possible self-occlusion or other occlusion in a single view. A suitable solution could be established by leveraging prior knowledge of the object shape. Numerous efforts have been devoted to combining the cognitive abilities of active users with the accuracy and efficiency of computational algorithms. However, most of these methods focus on modeling limited types of shapes with regular geometry (e.g., cuboids, swept surfaces, extruded objects, etc.) or a specific shape class (e.g., architecture, garments, trees, etc.). In this work, we present a novel interactive imageguided modeling approach which greatly extends the capability of existing methods. To model more complicated shapes given a single image, our approach is closely related to the lofting technique used for creating freeform surfaces in CAD systems. Our method is also based on the observation that most man-made or natural objects have irregular 3D shapes but rather regular 2D cross sections. By exploiting the regularities in these sections as well as additional geometric constraints, we can model a large variety of objects from a single image with minimal user input. Such lofted shapes can be readily found in
Abstract Modeling the complete geometry of general shapes from a single image is an ill-posed problem. User hints are often incorporated to resolve ambiguities and provide guidance during the modeling process. In this work, we present a novel interactive approach for extracting high-quality freeform shapes from a single image. This is inspired by the popular lofting technique in many CAD systems, and only requires minimal user input. Given an input image, the user only needs to sketch several projected cross sections, provide a “main axis”, and specify some geometric relations. Our algorithm then automatically optimizes the common normal to the sections with respect to these constraints, and interpolates between the sections, resulting in a high-quality 3D model that conforms to both the original image and the user input. The entire modeling session is efficient and intuitive. We demonstrate the effectiveness of our approach based on qualitative tests on a vari
Data Loading...
