|
|
|
|
| Bachelorprojekt - Softwareteknologi | Project No. 0064: Semiautomatic Modelling of 3D Objects |
| Aktuelle | Tidligere |
3D content creation (making 3D models of characters or objects) is a manual and time consuming process, which is a big obstacle in the way of even more widespread use of 3D computer graphics. Lots of user generated content is found online, but it is mostly text, images or videos - few 3D models are available although it is possible to find examples.
One way to make it easy for average users to produce 3D content is via semiautomatic modelling from photos. Take some pictures of your favorite car and quickly make a 3D model. This is already possible to some extent - e.g. using Google Sketchup - but existing systems leave most of the modelling to the human being while providing tools that simplty assist the user in the modelling. In the example of Google Sketchup, the user is assisted in laying down the coordinate system for 3D modelling so that it matches the object in a set of photographs.
This project is about taking the next step in this direction. Basically, we want the students to make a simple tool for template based 3D modelling where the user: segments the image into foreground and background. Coarsely aligns a template to the object (a box or a cylinder in most cases). and the computer then fits the template to the object.
To make the task feasible in a bachelor project, we will focus on simple objects and provide you with code for some of the hard computer vision tasks.
From a computer science point of view, the challenges are: - Coding a front end (UI design and implementation) - Interfacing to APIs for computer graphics and geometry processing. - Implementation of vision and graphics algorithms.
| Prerequisites: | Preferably some knowledge of computer graphics and computer vision. |
Supervisor(s) Andreas Bærentzen, Henrik Aanæs
| Sidst opdateret: Oct 31, 2011 af Hans Henrik Løvengreen |