Reference-guided exposure fusion in dynamic scenes

► We extend exposure fusion to dynamic scenes. ► A new quality assessment system is developed to account for ghosting artifacts. ► The proposed method is efficient and can produce different types of fusion results. ► The proposed method can be applied to flash and no-flash photography. Unlike high d...

Full description

Saved in:
Bibliographic Details
Published in:Journal of visual communication and image representation 2012-04, Vol.23 (3), p.467-475
Main Authors: Zhang, Wei, Cham, Wai-Kuen
Format: Article
Language:English
Subjects:
Consistency
Deghosting
Dynamical systems
Dynamics
Exposure fusion
Flash and no-flash photography
Ghosting
High dynamic range imaging
Image gradient
Quality assessment
Tone mapping
Weighting map
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:► We extend exposure fusion to dynamic scenes. ► A new quality assessment system is developed to account for ghosting artifacts. ► The proposed method is efficient and can produce different types of fusion results. ► The proposed method can be applied to flash and no-flash photography. Unlike high dynamic range (HDR) imaging, exposure fusion is a process of generating a tonemapped-like HDR image directly by fusing a series of bracketed images. Since it frees users from the tedious radiometric calibration and tone mapping steps, this technique is getting more and more popular, and becomes a basic tool in many graphics software. The main drawback of exposure fusion is its limitation to static scenes and any object movement of the target scene will incur severe ghosting artifacts in the fused result. In this paper, we intend to overcome this limitation and make exposure fusion applicable in dynamic scenes. A new quality assessment system is developed, where both temporal consistency and spatial consistency are introduced to account for ghosting artifacts. Experimental results of various dynamic scenes are shown to prove the effectiveness of the proposed method.
ISSN:1047-3203
1095-9076
DOI:10.1016/j.jvcir.2012.01.006