QUALITY MEASUREMENT FOR RECONSTRUCTED RGB IMAGE VIA NOISY ENVIRONMENTS

Abstract

Image compression and decompression process could be quite affected by noisy environment during transmitting/receiving medium. This paper develops a procedure which finds the effect of noisy environments on the reconstructed RGB images. The image has been degraded by three kinds of noises then; noisy image planes are transformed into new domain of four bands by applying the first level 2D DWT. The inverse 2D DWT is applied on the noisy RGB planes to reconstruct; concatenate; and restore the original transmitted image. The quality of re-stored images is measured by applying SNR/PSNR with respect to the noise variances. The SNR/PSNR dB curves are used for comparing different noisy environment effects on the quality of reconstructed RGB images. The paper provides basic procedure for calculating scale factors used for reconstructing images directly in SNR/PSNR units. The SNR/PSNR dB curves for Gray images satisfied better result than RGB for all testing conditions, while speckle noise was relatively the most stable degrading noise that had maximum dB values over wide noise variance. Salt & pepper noise had the worst dB curves among Gaussian and speckle. The intersection points of the dB curves at 0.5 density noise is discussed and concluded to find out the SNR/PSNR behavior at this degradation value.