Using Bayesian Networks with Hidden Nodes to Recognise Neural Cell Morphology

Bayesian decision trees are based on a formal assumption that the unconnected nodes are conditionally independent given the states of their parent nodes. This assumption does not necessarily hold in practice and may lead to loss of accuracy. We propose a methodology whereby naive Bayesian networks are adapted by the addition of hidden nodes to model the data dependencies more accurately. We examined the methodology in a computer vision application to classify and count the neural cell automatically. Our results show that a modified network with two hidden nodes achieved significantly better performance with an average prediction accuracy of 83.9% compared to 59.31% achieved by the original network.