Synthesizing new views from a pair of stereo images

T. Y. Chao*, Hsueh-Ming Hang, Sheng-Jyh Wang

*Corresponding author for this work

Research output: Contribution to journalConference articlepeer-review


Our goal in this study is to construct a 3-D model from a pair of (stereo) images and then project this 3-D model to image planes at new locations and orientations. We first compute the disparity map from a pair of stereo images. Although the disparity map may contain defects, we can still calculate the depth map of the entire scene by filling in the missing (occluded) pixels using bilinear interpolation. One step further, we synthesize new stereo images at different camera locations using the 3-D information obtained from the given stereo pair. The disparity map used to generate depth information is one key step in constructing 3-D scenes. Therefore, in this paper we investigate various types of occlusion to help analyzing the disparity map errors and methods that can provide consistent disparity estimates. The edge-directed Modified Dynamic Programming scheme with Adaptive Window (MDPAW), which significant improves the disparity map estimates, is thus proposed. Our preliminary simulations show quite promising results.

Original languageEnglish
Pages (from-to)87-98
Number of pages12
JournalProceedings of SPIE - The International Society for Optical Engineering
StatePublished - 24 Jan 1998
EventStereoscopic Displays and Virtual Reality Systems V - San Jose, CA, United States
Duration: 26 Jan 199829 Jan 1998


  • 2D/3D conversion
  • 3D modeling
  • Adaptive window
  • Correspondence
  • Depth estimation
  • Disparity estimation
  • Dynamic programming
  • Matching indices
  • Occlusion
  • Stereo

Fingerprint Dive into the research topics of 'Synthesizing new views from a pair of stereo images'. Together they form a unique fingerprint.

Cite this