Restoration of blur and noisy images using dempster-belief probabilistic approximation
Deepa Bamne, Ramratan Ahirwal, Yogendra Kumar Jain · International Journal of Management, IT, and Engineering · 2015
Digital images are very widely used for many scientific and forensic investigations but the quality of digital image snapshots may be degraded. Digital snapshots may gets blur due to bad focus of camera, relative motion between the scene to be capture and camera etc. This work proposes a probabilistic approach to recover these images in order to improve the quality of degraded image. A Dempster-Belief probabilistic approximation is used to approximate the blur distribution (PSF). Wiener filter uses this approximated PSF in order to get the noiseless and unblur image. Also padding is done to recover the image after proper blur approximation. This work also makes use of Haar wavelet transform. The proposed method gives the better result from the previous method. It seems to be that the PSNR and Mean Square Rate is better in the proposed work.