A Variational Bayesian Inference Framework for Multiview Depth Image Enhancement
2012 (English)In: Proceedings - 2012 IEEE International Symposium on Multimedia, ISM 2012, IEEE , 2012, 183-190 p.Conference paper (Refereed)
In this paper, a general model-based framework for multiview depth image enhancement is proposed. Depth imagery plays a pivotal role in emerging free-viewpoint television. This technology requires high quality virtual view synthesis to enable viewers to move freely in a dynamic real world scene. Depth imagery of different viewpoints is used to synthesize an arbitrary number of novel views. Usually, the depth imagery is estimated individually by stereo-matching algorithms and, hence, shows lack of inter-view consistency. This inconsistency affects the quality of view synthesis negatively. This paper enhances the inter-view consistency of multiview depth imagery by using a variational Bayesian inference framework. First, our approach classifies the color information in the multiview color imagery. Second, using the resulting color clusters, we classify the corresponding depth values in the multiview depth imagery. Each clustered depth image is subject to further subclustering. Finally, the resulting mean of the sub-clusters is used to enhance the depth imagery at multiple viewpoints. Experiments show that our approach improves the quality of virtual views by up to 0.25 dB.
Place, publisher, year, edition, pages
IEEE , 2012. 183-190 p.
Depth enhancement, Gaussian mixture model, Multiview video, Variational Bayesian inference
Engineering and Technology
IdentifiersURN: urn:nbn:se:kth:diva-102704DOI: 10.1109/ISM.2012.44ISI: 000317430600036ScopusID: 2-s2.0-84874244618ISBN: 978-076954875-3OAI: oai:DiVA.org:kth-102704DiVA: diva2:555934
14th IEEE International Symposium on Multimedia, ISM 2012;Irvine, CA;10 December 2012 through 12 December 2012
FunderICT - The Next Generation
QC 201303082012-09-222012-09-222013-05-23Bibliographically approved