University of Illinois Urbana-Champaign

Intermediate view generation by stereo matching

Chen, Yuhao

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/104002

Description

Title
Intermediate view generation by stereo matching
Author(s)
Chen, Yuhao
Contributor(s)
Baryshnikov, Yuliy
Issue Date
2019-05
Keyword(s)
  • stereo
  • dynamic programming
  • Laplacian filter
  • dynamic time warping
Abstract
Previous stereo matching algorithms have been developed in order to construct the accurate disparity map by using the Epipolar geometry and Rectification and finding the stereo matching on a per-pixel basis. In this paper, our final goal is to generate an intermediate view and we only focus on pictures with linear patterns. We use image resizing and Laplacian filters to find the patterns and then use segmentation-based techniques to match the patterns. After that, we create a new algorithm based on dynamic programming to find the global optimization for each segment of the picture. The approach works by computing the minimum-cost path through the matrix of all pairwise matching costs between two corresponding scanlines. In the algorithm, we defined two different cost functions for different situations. One is the total sum of the differences of degrees between each adjacent segment of lines in the matrix. The other one is the distance of current alignment to the previous one, where the distance is found by Dynamic Time Warping. In comparing the results, our algorithm should be much faster on the scenes with linear patterns. In the future, we hope we can improve the runtime of Dynamic Time Warping and develop the algorithm to all different kinds of patterns.
Type of Resource
text
Language
en
Permalink
http://hdl.handle.net/2142/104002

Owning Collections

Senior Theses - Electrical and Computer Engineering PRIMARY
The best of ECE undergraduate research