TY  - JOUR
T1  - Basic Methods for Motion Detection in Images Sequence
AU - , T. Bouden AU - , N. Doghmane 
JO  - Asian Journal of Information Technology
VL  - 6
IS  - 3
SP  - 296
EP  - 302
PY  - 2007
DA  - 2001/08/19
SN  - 1682-3915
DO  - ajit.2007.296.302
UR  - https://makhillpublications.co/view-article.php?doi=ajit.2007.296.302
KW  - Images sequence
KW  -segmentation
KW  -detection
KW  -moving objects
KW  -Markov model
KW  -maximum likelihood
KW  -site
KW  -clique
KW  -determinist and stochastic relaxation
AB  - In the physical world, motion segmentation of images sequences is based on visual motion perception. This does not depend on prior interpretation or recognition of shape and form. However, it does depend  on motion information (spatiotemporal object-environment relations). It is generally recognized that the analysis of moving objects proceeds in four stages: The first is the detection of variations in intensity over time in the environment. The second is the segmentation of moving areas and objects masks building. The third is the estimation of motion parameters. The fourth one is the 3D motion interpretation. In the study, we are dealing with detection and region-based segmentation methods. These methods may easily extend to estimate motion parameters. Here we are mainly concerned with comparing studies using determinist and stochastic modelling (images difference, maximum likelihood detector and Markov random field model) to detect the moving objects masks.
ER  -