DocumentCode
598041
Title
Adaptive nonlocal range filter in depth map coding
Author
Ilsoon Lim ; Jaejoon Lee
fYear
2012
fDate
Sept. 30 2012-Oct. 3 2012
Firstpage
1285
Lastpage
1288
Abstract
In this paper, we present a novel depth map coding scheme to improve coding efficiency. For this purpose, we develop an adaptive nonlocal range filter that restores the coded depth map by the noise suppression. Unlike the previous works that adapted the original nonlocal means algorithm or the original bilateral filter to the depth data, our approach is fundamentally different in that the nonlocal principle in the color data and the range filter in the depth map are taken into account, respectively. Furthermore, the optimal fall-off parameters that control the smoothing strength are determined at the encoder and the restoration is performed with the optimal fall-off parameter at the decoder. This technique allows for the strong slope restoration at the boundary and the noise smoothing near the boundary. Finally, the depth map coding that performs the new approach is implemented and this results in 9.71%(rate-distortion evaluation) and 8.71%(rate-quality evaluation) BD-bitrate reduction on average when compared with 3DV HEVC.
Keywords
adaptive filters; codecs; maximum likelihood decoding; smoothing methods; video coding; 3DV HEVC; BD-bitrate reduction; adaptive nonlocal range filter; bilateral filter; coding efficiency; color data; depth map coding; encoder; fall-off parameter; noise smoothing; noise suppression; optimal fall-off parameters; range filter; rate-distortion evaluation; rate-quality evaluation; smoothing strength; Bit rate; Decoding; Encoding; Image color analysis; Kernel; Noise; Noise reduction; Range filter; depth map restoration; noise removal; nonlocal means;
fLanguage
English
Publisher
ieee
Conference_Titel
Image Processing (ICIP), 2012 19th IEEE International Conference on
Conference_Location
Orlando, FL
ISSN
1522-4880
Print_ISBN
978-1-4673-2534-9
Electronic_ISBN
1522-4880
Type
conf
DOI
10.1109/ICIP.2012.6467102
Filename
6467102
Link To Document