Automatic 2D to 3D video and image conversion based on global depth map
Authors:
- 1 December 2015
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
Abstract
3d technology brings a new era of entertainment to the human race. It offered a wide array of possibilities in near future in almost every walk of life and in entertainment segment. 3D content generation is the important step in 3D systems. Special cameras such as stereoscopic dual camera, depth range camera etc. are designed to generate the 3D model of a scene directly. There are different techniques to generate the 3D content. But the problem is our current and past media data are in 2D which needs to convert into 3D. This is where the importance of 2D to 3D transformation arises. In this paper proposed real time 3D image and video creation by depth map estimation. Depth map estimation can be done in two methods. One is based on depth fusion method and other is based on saliency map of an image. In dataset image estimate the depth map from depth fusion method and then depth is refined by color spatial variance. In non dataset images we find depth map by global saliency method. Experimental result demonstrates that the proposed technique convey better performance compared to the state-of-the-art of methods.Keywords
This publication has 7 references indexed in Scilit:
- Example-Based Video Stereolization With Foreground Segmentation and Depth PropagationIEEE Transactions on Multimedia, 2014
- Learning-Based, Automatic 2D-to-3D Image and Video ConversionIEEE Transactions on Image Processing, 2013
- An automatic 2D to 3D conversion algorithm using multi-depth cues2012
- Video Stereolization: Combining Motion Analysis with User Interaction2011
- Semi-automatic stereo extraction from video footage2009
- Stereoscopic video generation based on efficient layered structure and motion estimation from a monoscopic image sequenceIEEE Transactions on Circuits and Systems for Video Technology, 2005
- Generation of stereoscopic image sequences using structure and rigid motion estimation by extended Kalman filters2003