Image segmentation based on variance feature

Abstract EN

Image segmentation, Which performs the noise filtering and edge preserving smoothing processes synchronously, are developed in this paper.

This process, cleans the noise, and in the same time, does not blur the edges of the image.

An ensemble of pixels, coupled in a local region, is used for grouping purpose.

This task is implemented via the introducing a set of neighborhood.

Such task may generate groups associated with a specific pixel based on the similarity principle.

The task of this work presented in this paper applies on the syntactic, medical, satellite images.

The simulation results demonstrate the effectiveness of the developed algorithm.