A unified approach for image segmentation using exact statistics

https://doi.org/10.1016/S0734-189X(89)80038-6Get rights and content

In this paper the concept of segmentation of anoisy digitized image is developed. The problem is treated as a problem of classification into classes comprising different types of edges and regions. A framework to represent these classes is established so that a unified approach for edge detection and region growing can be considered simultaneously. A two-stage algorithm is developed in this framework. The first stage classifies the pixels on a qualitative basis and the second stage classifies on a quantitative basis. An exact statistic to perform the requisite test in the second stage is derived. Two illustrations of this algorithm are presented.

References (15)

There are more references available in the full text version of this article.

Cited by (9)

  • A MRF model-based segmentation approach to classification for multispectral imagery

    2002, IEEE Transactions on Geoscience and Remote Sensing
View all citing articles on Scopus
View full text