Image restoration using a hybrid approach based on DWT and SMKF

Various approaches based on Kalman filtering exist in the literature for image restoration. The reduced order model Kalman filter (ROMKF) has comparable performance with less computational complexity. For further reduction in computational complexity, we suggest a simplified model Kalman filter (SMKF) for image restoration. Furthermore, a hybrid approach based on discrete wavelet transform (DWT) and SMKF is proposed for image restoration with better SNRs especially when the observed image signal-to-noise ratio is low. In the first step, the approach uses the DWT with few resolution levels and a moderate threshold value for denoising the image. The denoised image will provide a better data to SMKF in the second step. The proposed approach is implemented on a visual image to evaluate its performance in comparison with the SMKF and DWT approaches