A New Textured Image Segmentation Algorithm Based on Gaussian Mixture Models

YU Peng,FENG Jufu,TONG Xingwei
DOI: https://doi.org/10.3969/j.issn.1671-8860.2005.06.011
2005-01-01
Abstract:This paper proposes a new textured image segmentation algorithm which integrates wavelet transform and Gaussian mixture models. This algorithm includes pretreatment, wavelet transform, training, calculating posterior probability and labeling image. Some aspects are analyzed in order to use Gaussian mixture models in textured image segmentation problem. Experiments show that the algorithm can give an excellent segmentation result.
What problem does this paper attempt to address?