Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 47605 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Kasaei, Shohreh
- Abstract:
- Nowadays TV broadcasting of soccer matches has attracted a lot of viewers. With advancement of technology, the broadcasting is also going under changes. One of these changes is using free-viewpoint video. In free-viewpoint video users can interactively choose any view to watch the match from. Generally free-viewpoint problem is defined as follows. We have sequential frames of a soccer match video from different views as input. We want to create sequential frames of the same video from a different view as output. There are several methods for solving this problem. We can classify these methods into to general approaches. Image-based and geometry-based. We will examine these approaches in chapter two of this thesis. In this research two methods are used to solve this problem. In the first one, the visual hull of a soccer player is created. This method is used to evaluate the results of the main method. It also can be used as the input of it. In the second method 3d model is created with optimization technics. An cost function on pixel’s depths will be defined. This function will be optimized with -expansion algorithm and optimized depth images will be yielded. At last the 3d model will be created from the computed depth images. In this research, a novel cost function is defined, so that better depth images are yielded. This cost function has four terms including photo consistency, silhouette consistency, depth smoothness, and pose consistency. This method will achieve good results when we have inaccurate calibration. Also when we have few input views,better results are achieved compared to previous methods. We will compare the results of these methods with two indicators in this research. The percentage of reconstruction and the accuracy of it. The percentage of reconstruction is the percentage of pixels in the ground truth silhouette image that are reconstructed in the new image. The accuracy of reconstruction is the percentage of pixels that are reconstructed correctly.The introduced method will achieve 6.09 % advantage in percentage of reconstruction and 0.73 % advantage in accuracy, compared to the previous method. In comparison to visual hull it will fall back by 10.38 % in percentage of reconstruction, but will lead by 13.32 % in accuracy. The runtime performance of the introduced method is 13.32% slower than the previous method
- Keywords:
- Three Dimensional Reconstruction ; Graph Cut ; Free Viewpoint Video ; Soccer Math ; Video Images ; Visual Holl Planes
Digital Object List
محتواي کتاب
Bookmark