A Spatially Constrained Generative Model and an EM Algorithm for Image Segmentation

• Published in: IEEE transactions on neural networks Vol. 18; no. 3; pp. 798 - 808
• Main Authors: Diplaros, A., Vlassis, N., Gevers, T.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.05.2007; Institute of Electrical and Electronics Engineers
• Subjects: Algorithms; Applied sciences; Artificial Intelligence; Bound optimization; Clustering algorithms; Computer science; control theory; systems; Computer Simulation; Connectionism. Neural networks; Constraints; Exact sciences and technology; expectation-maximization (EM) algorithm; Hidden Markov models; hidden Markov random fields (MRFs); Image color analysis; Image edge detection; Image Enhancement - methods; Image Interpretation, Computer-Assisted - methods; Image segmentation; Informatics; Intelligent sensors; Intelligent systems; Iterative algorithms; Labels; Likelihood Functions; Mathematical models; Models, Statistical; Pattern Recognition, Automated - methods; Pattern recognition. Digital image processing. Computational geometry; Pixel; Pixels; spatial clustering; Cluster analysis; Mixed distribution; Markov process; expectation-maximization (EM) algorithm; Lower bound; hidden Markov random fields (MRFs); Probabilistic approach; Similarity; Image processing; Prior distribution; spatial clustering; Neural network; Modeling; Competitive algorithms; Optimization; Image segmentation; Gaussian process; Classification; Hidden Markov model; Log file; EM algorithm; Bound optimization
• ISSN: 1045-9227; 1941-0093
• DOI: 10.1109/TNN.2007.891190

In this paper, we present a novel spatially constrained generative model and an expectation-maximization (EM) algorithm for model-based image segmentation. The generative model assumes that the unobserved class labels of neighboring pixels in the image are generated by prior distributions with similar parameters, where similarity is defined by entropic quantities relating to the neighboring priors. In order to estimate model parameters from observations, we derive a spatially constrained EM algorithm that iteratively maximizes a lower bound on the data log-likelihood, where the penalty term is data-dependent. Our algorithm is very easy to implement and is similar to the standard EM algorithm for Gaussian mixtures with the main difference that the labels posteriors are "smoothed" over pixels between each E- and M-step by a standard image filter. Experiments on synthetic and real images show that our algorithm achieves competitive segmentation results compared to other Markov-based methods, and is in general faster