Sciences & technologie. B, Sciences de l’ingénieur
Volume 0, Numéro 20, Pages 11-19
2003-12-31
Authors : Benmoussat N . Belbachir M.f .
In this paper, we propose a new technique for motion compensated filtering of noisy image sequences which aims to achieve an efficient noise reduction without introducing blurring artefacts. The technique consists in a cascade of the versions of the temporal and the spatiotemporal Adaptive Weighted Averaging (AWA) filter proposed in [11]. We referred the technique as a Two-Stage AWA filter (TS-AWA) [1][2]. The process consists in: first, the noisy observation is filtered temporally using the temporal AWA filter (AWA 1-D) along a large motion compensated temporal support. Second, the resulting image sequence is then processed by the spatiotemporal AWA filter (AWA 3- D) along a smaller support which the motion trajectories are estimated after the first stage. In each stage, motion trajectories are estimated with a modified block matching motion estimation algorithm [3]. Simulation results show that the proposed TS-AWA filter outperforms the AWA 3-D filter in both terms of visual quality and Signal to Noise Ratio improvement.
Image sequence processing, motion estimation, motion compensation, adaptive filtering.
A. Abbad
.
A. Belaidi .
.
pages 20-25.
A. Bounoua
.
S.m.a Wady
.
M. Gharbi
.
pages 90-96.
أيت عبد القادر عبد العزيز
.
ص 235-248.
Maouche Née Ketfi Salima
.
pages 9-17.