A New Perceptual Quality Assessment for Image Coding

M.A. Ben Ayed, A. Samet, M. Loulou, and N. Masmoudi (Tunisia)


image quality assessment, human visualsystem (HVS), mean squared error (MSE), weightedmean squared error (WMSE).


Compression of digital images has been studied for many years. However, algorithms have been designed and optimized based on simple error measures like peak signal-to-noise ratio (PSNR) despite the fact that such measures correlate poorly with perceived quality. The problem is that PSNR does not take into account viewing conditions and the comportment of the human visual system (HVS). These problems necessitate methods of objective quality assessment. Ideally, such a system would perceive and measure impairments just like a human being. This paper presents a quality metric for image: weighted mean squared error.

Important Links:

Go Back