Color Image Segmentation Based On Principal Component Analysis With Application of Firefly Algorithm And Gaussian Mixture Model

. International Journal of Image Processing (IJIP), 12 (4): 101 - 112 (December 2018)


In this paper we propose a segmentation method for multi-spectral images in the HSV space, based on the Principal Component Analysis to generate grayscale images. Then the Firefly Algorithm has been applied on the gray-level images in a histogram-based research of cluster centroids. The FA is a metaheuristic optimization algorithm, centered on the flashing behaviour of fireflies. The Firefly Algorithm is performed to determine automatically the number of clusters and to select the gray levels for partitioning pixels into homogeneous regions. Successively, these gray values are employed during the initialization step of a Gaussian Mixture Model for estimation of parameters, evaluated through the Expectation-Maximization technique. The coefficients of the linear super-position of Gaussians can be regarded as the prior probabilities of each component. Applying the Bayes rule, the posterior probabilities have been estimated and their maxima are used to assign each pixel to the clusters, according to their gray values.

Links and resources