Development of improved SSIM quality index for compressed medical images

This paper presents the development of improved structural similarity index (SSIM) quality assessment parameter for compressed medical images. The SSIM has been proved to be better objective quality assessment metric which exploits the structural similarity in the viewing field. In this paper, the SSIM quality index has been modified by achieving optimal value of arbitrary constant K used in SSIM index expression. The analysis has been done for three different types of medical image; MRI scan, CT scan and ultrasound image. SSIM index along with PSNR, MSE and mean-opinion-score (MOS) are computed for SPIHT compressed medical images at varying compression rates by setting values of K in the range 0.02 to 2.0. The value of K giving maximum correlation coefficient (CC) between SSIM index and MOS is selected as optimal value. It is found that optimal values of K equal to 0.5, 0.05 and 0.1 for MRI, CT scan and ultrasound images respectively.