Low-complexity stereo matching and viewpoint interpolation in embedded consumer applications

Lu Zhang, Ke Zhang, Jiangbo Lu, Tian-Sheuan Chang, Gauthier Lafruit

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

Abstract

Viewpoint interpolation is the process of synthesizing plausible in-between views - so-called virtual camera views - from a couple of surrounding fixed camera views. To make viewpoint interpolation possible for low/moderate-power consumer applications, a further quality/complexity trade-off study is required to conciliate algorithmic quality to architectural performance. In essence, the inter-dependencies between the different algorithmic steps in the processing chain are thoroughly analyzed, aiming at an overall quality-performance model that pinpoints which algorithmic functionalities can be simplified with minor global input-output quality degradation, while maximally reducing their implementation complexity w.r.t. arithmetic and line buffer requirements. Compared to state-of-the-art CPU and GPU platforms running at several GHz clock speed, our low-power 100 MHz FPGA implementation achieves speedups with one to two orders of magnitude, without impeding on the visual quality, reaching over 100 frames per second VGA high-quality, 64-disparity search range stereo matching and enabling viewpoint interpolation in low-power, embedded applications.

Original languageEnglish
Title of host publicationDepth Map and 3D Imaging Applications
Subtitle of host publicationAlgorithms and Technologies
PublisherIGI Global
Pages307-330
Number of pages24
ISBN (Print)9781613503263
DOIs
StatePublished - 1 Dec 2011

Fingerprint Dive into the research topics of 'Low-complexity stereo matching and viewpoint interpolation in embedded consumer applications'. Together they form a unique fingerprint.

Cite this