Hierarchical hidden markov models in image segmentation

Hidden Markov Models have been extensively used in various fields, especially in speech recognition, biology, image and signal processing and digital communication. They are well known by their effectivenss in modeling the correlations between adjacent symbols, domains or events, but they often suffer from high dimensionality problems. In this work, we propose two approaches to reduce the execution time of Hidden Markov Chain with Independent Noise used in image segmentation. The first one consists of dividing the image into blocks, each of them is treated independently of other. In the second approach, we have divided the observations into blocks, but the treatment of each block depends on its previous one. The obtained results, show that our approaches outperform standard one, and contribute efficiently to reduce the execution time and the number of iterations ensuring the convergence. © 2020 National Research Nuclear University. All rights reserved.