In this paper, we propose a novel technique for modelbased recognition of complex object motion trajectories using Gaussian Mixture Models (GMM). We build our models on Principal ...
Faisal I. Bashir, Ashfaq A. Khokhar, Dan Schonfeld
Exploiting prior knowledge, we use Bayesian estimation to localize a source heard by a fixed sensor network. The method has two main aspects: Firstly, the probability density fun...
This paper proposes a novel Bayesian approximation inference method for mixture modeling. Our key idea is to factorize marginal log-likelihood using a variational distribution ove...
Background: In addition to their use in detecting undesired real-time PCR products, melting temperatures are useful for detecting variations in the desired target sequences. Metho...
In this paper, a new theoretical framework based on hidden Markov model (HMM) and independent component analysis (ICA) mixture model is presented for content analysis of video, nam...