Unsupervised textured image segmentation

George K. Gregoriou, Oleh J. Tretiak

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


A new algorithm for unsupervised textured image segmentation is presented. The image is comprised of M textured regions, each of which is modeled by a stationary Gaussian Markov random field. A feature vector is computed for each pixel in the original image where these vectors are normally distributed and cluster about some vector means. Thus, the problem is reduced to one of restoring a vectorvalued underlying field embedded in additive Gaussian noise. The vector means corresponding to the different regions are estimated by using the EM algorithm. By using these means an iterative algorithm is employed, where the underlying field is modeled as a multi-level logistic Markov random field. The results obtained on 2-region and 4-region textured images are impressive and the classification error is less than 3%. The algorithm is not limited to textured images but can also be applied to any vector-valued signals.

Original languageEnglish
Title of host publicationICASSP 1992 - 1992 International Conference on Acoustics, Speech, and Signal Processing
PublisherInstitute of Electrical and Electronics Engineers Inc.
Number of pages4
ISBN (Electronic)0780305329
Publication statusPublished - 1992
Event1992 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 1992 - San Francisco, United States
Duration: 23 Mar 199226 Mar 1992


Other1992 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 1992
Country/TerritoryUnited States
CitySan Francisco


Dive into the research topics of 'Unsupervised textured image segmentation'. Together they form a unique fingerprint.

Cite this