Abstract--We present a real-time incremental approach to motion segmentation operating on sparse feature points. In contrast to previous work, the algorithm allows for a variable n...
We propose in this paper a novel approach to the induction of the structure of Hidden Markov Models. The induced model is seen as a lumped process of a Markov chain. It is construc...
We present a tool for real-time visualization of motion features in 2D image sequences. The motion is estimated through an eigenvector analysis of the spatiotemporal structure ten...
Many computer vision problems can be formulated in
a Bayesian framework with Markov Random Field (MRF)
or Conditional Random Field (CRF) priors. Usually, the
model assumes that ...
With the shrinking of technology feature sizes, the share of leakage in total power consumption of digital systems continues to grow. Traditional dynamic voltage scaling (DVS) fail...