RESTORATION OF COLOR IMAGES BY SEPERATING GAUSSIAN NOISE USING ADAPTIVE HISTOGRAM EQUALIZATION AND GAUSSIAN FILTERING

Abstract

Image Restoration is an important phenomenon in digital image processing that deals with restoring an
image which is as sharp and clear as the original image.Image may suffer deterioration as a result of pixel missing
and noise.In order to obtain recovered image very close to the original one from corrupted image employing a
mathematical degradation and restoration model. In this paper we are using jpeg images as input and then resizing it
to size of 512*512 as we are performing our work on the images of this size only.Later we have performed ROI
(Reason of Interest) operation.Then we've applied numerous deterioration levels ,added Gaussian noise and used
Gaussian Filtration method along with Adaptive Histogram Equalization for restoration of deteriorated image.
We've calculated PSNR (Peak Signal to Noise Ratio) and MSE (Mean sq. Error) so as to calculate the
performance.We are getting PSNR value for Lena as 38.77db for 25%, 38.24db for 50% and 37.90db for 75%
deterioration at noise=5% while at noise =15%,our proposed method calculated PSNR as 36.97db for 25%,36.22db
for 50% and 36.05db for 75% deterioration of image.