128
click to vote
IJCV
2008
15 years 3 months ago
2008
We present an unsupervised approach for learning a generative layered representation of a scene from a video for motion segmentation. The learnt model is a composition of layers, ...