It is very challenging to generate a high dynamic range (HDR)-like low dynamic range (LDR) image, called wide dynamic range (WDR) image, without the ghost artifact from multiple differently exposed LDR images for dynamic scenes. Since the ghost artifact is caused by the moving objects in the input LDR images, it is required to properly detect the motion region to prevent the ghost artifact in the WDR image. In order to generate the ghost-free WDR image, this paper presents a method which can effectively detect and process the motion region by using the structural similarity in the co-located local region of multiple LDR images. Experimental result shows that the proposed method provides superior performance in terms of the extent of ghost artifact as compared with the existing methods.
[1]
Jan Kautz,et al.
Exposure Fusion
,
2009,
15th Pacific Conference on Computer Graphics and Applications (PG'07).
[2]
Sung-Jea Ko,et al.
Eigen directional bit-planes for robust face recognition
,
2014,
IEEE Transactions on Consumer Electronics.
[3]
Shutao Li,et al.
Fast multi-exposure image fusion with median filter and recursive filter
,
2012,
IEEE Transactions on Consumer Electronics.
[4]
Wai-kuen Cham,et al.
Gradient-Directed Multiexposure Composition
,
2012,
IEEE Transactions on Image Processing.
[5]
Jan Kautz,et al.
Bitmap Movement Detection: HDR for Dynamic Scenes
,
2010,
2010 Conference on Visual Media Production.