Motion segmentation using Markov random field model for accurate moving object segmentation
Proceedings of the 2nd international conference on Ubiquitous information management and communication
Moving object segmentation by background subtraction and temporal analysis
Image and Vision Computing
Simultaneous motion estimation and segmentation
IEEE Transactions on Image Processing
An efficient two-pass MAP-MRF algorithm for motion estimation based on mean field theory
IEEE Transactions on Circuits and Systems for Video Technology
| Hi-index | 0.00 |
This paper provides a new motion segmentation algorithm in image sequences based on gamma distribution. Conventional methods use a Gaussian mixture model (GMM) for motion segmentation. They also assume that the number of probability density function (PDF) of velocity vector's magnitude or pixel difference values is two. Therefore, they have poor performance in motion segmentation when the number of PDF is more than three. We propose a new and accurate motion segmentation method based on the gamma distribution of the velocity vector's magnitude. The proposed motion segmentation algorithm consists of pixel labeling and motion segmentation steps. In the pixel labeling step, we assign a label to each pixel according to the magnitude of velocity vector by optical flow analysis. In the motion segmentation step, we use energy minimization method based on a Markov random field (MRF) for noise reduction. Experimental results show that our proposed method can provide fine motion segmentation results compared with the conventional methods.