Pregled bibliografske jedinice broj: 560899
Adaptive motion-estimation-mode selection for depth video coding
Adaptive motion-estimation-mode selection for depth video coding // Proceedings ICASSP 2010 / Hansen, John (ur.).
Dallas (TX): Institute of Electrical and Electronics Engineers (IEEE), 2010. str. 702-705 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 560899 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Adaptive motion-estimation-mode selection for depth video coding
Autori
Kamolrat, B. ; Fernando, W.A.C. ; Mrak, Marta
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
Proceedings ICASSP 2010
/ Hansen, John - Dallas (TX) : Institute of Electrical and Electronics Engineers (IEEE), 2010, 702-705
ISBN
978-1-4244-4295-9
Skup
2010 IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP)
Mjesto i datum
Dallas (TX), Sjedinjene Američke Države, 14.03.2010. - 19.03.2010
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
3D video; block matching; depth imagebased rendering (DIBR); motion estimation
Sažetak
An effective representation of 3D video in future 3D-TV systems consists of monoscopic video (colour component) and associated per-pixel depth information (depth component). As depth component indicates relative distance between objects within the scene and a camera, pixel values change not only when objects move in vertical and horizontal directions but also when they move in a depth direction. Instead of predicting motion of objects in two directions as appearing in traditional video codecs, three-dimensional block matching (3D-BM) achieves more accurate motion estimation in depth video coding. However overall performance of 3D-BM exceeds that of the traditional two-dimensional block matching (2D-BM) only at high bit rate. In this paper, an adaptive 2D-3D BM selection algorithm is introduced to compromise performance of 2DBM and 3D-BM. The Lagrangian optimisation algorithm is applied to select motion estimation mode at a block level. The experiment results reveal that the proposed adaptive motion-estimation-mode selection can improve the performance of 3D-BM at low bit rate while advantages of 3D-BM are preserved at high bit rate.
Izvorni jezik
Engleski
Znanstvena područja
Elektrotehnika, Računarstvo
POVEZANOST RADA
Projekti:
036-0982560-1643 - Inteligentno određivanje značajki slike u sustavima za otkrivanje znanja (Grgić, Mislav, MZO ) ( CroRIS)
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb
Profili:
Marta Mrak
(autor)