Real-time Stereoscopic Image Generation Using Depth Image Based Rendering with Virtual View Point Estimation
| Autor: | Hsuan-ChihChen, 陳炫志 |
|---|---|
| Rok vydání: | 2010 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 98 Three dimensional applications are widely spread all over the world. In this thesis, we proposed a stereoscopic image generation system using Depth Image Based Rendering (DIBR) method. Due to the unfavorable artifacts produced by the DIBR algorithm, researchers have developed various algorithms to handle the problem. The most common one is to smooth the depth map before rendering. However, pre-processing of the depth map usually generates other artifacts, such as rubber sheet artifact, geometry distortion and even the degradation of depth perception. In order to avoid these defects, we proposed a solution with another approach based on virtual view point estimation. We first analyze the feature of the holes produced by the DIBR method which tends to generate big holes in the regions with large variation of depth. We develop a virtual view estimation method similar to motion estimation to fill these holes, that is, to search the most similar block and fill the blank pixels by the corresponding pixels of the best candidate block. This simple and regular filling method is suitable for hardware implementation and is able to achieve real-time applications. The results show that our method works well on images with monotonous background and the strong depth perception of the main object in the scene can be preserved. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|