A performance comparison of two WS filters for image reconstruction technique under different image types

In this paper, we compare two weighted sum (WS) linear filters for image reconstruction technique under several image types. Two WS filters include hard partition-based weighted sum (HPWS) filter and subspace hard partition-based weighted sum (S-HPWS) filter. Four image types are considered including aerial image, face image, text image and sequence image. The simulation results show that the S-HPWS filter has PSNR more than the HPWS filter for face image type. For aerial image and sequence image, the S-HPWS filter has PSNR less than the HPWS filter at high noise level. However, simulation time of the S-HPWS filter is lower than the HPWS filter for all cases. On the other hand, when we compare both techniques under different image types, we found that aerial image is most appropriate to reconstruct by the HPWS filter. For face image is most appropriate to reconstruct by the S-HPWS filter. For text image get poor performance from both types of filter.