Performance analysis of noise removal techniques for digital images

Abstract EN

Noise has impairment effect on image so it destroys its information.

In order to overcome this problem, it is necessary to remove noise before extracting information and image processing as one of the preprocessing steps.

A brief overview of various sources and types of image noise as well as noise removal technique of digital images has been produced in this paper.

The general idea of this paper is to perform segmentation of the corrupted image before using noise removal filter.

Noise removal filter has been used to eliminate noise for each segment of image individually after segmentation process.

Then comparative study between noise removal filtering for whole image and segmentation based method has been produced with performance analysis in terms of MSE and PSNR.

In this paper, MSE and PSNR are calculated for the resultant reconstructed image compared with the original image.

Six image samples are used to test the Segmentation based noise removal approach.

The segmentation based technique has shown best results in terms of MSE and PSNR in comparison with traditional noise removal method based filter.