Image restoration with multiplicative noise: incorporating the sensor nonlinearity

A linear minimum mean-square-error deconvolution filter in the presence of multiplicative noise is derived. The importance of incorporating the nonlinear sensor characteristics into the restoration of noisy and blurred scanned photographic images is discussed. It is proposed to restore images in the 'exposure domain' where a linear convolutional relationship between the original and the observed images can be established. Results are presented on restoring photographic blurred images using the proposed filter in the exposure domain, whereas the use of the classical Wiener filter in the density domain, with additive noise assumption, does not yield any visible improvement.<
>