A previously published technique that combined a homomorphic transformation with a locally adaptive algorithm for image restoration in the presence of signal-dependent noise has been adapted for the case when the model for the. signal-dependent noise is unknown. The noise variance required for the homomorphic transformation is estimated from measurements over a small window, and the homomorphic transformation to make the noise additive is incorporated noise into the locally adaptive algorithm, resulting in a modified James–Stein algorithm valid for any signal- dependent noise.
© 1986 Optical Society of America
Original Manuscript: October 21, 1985
Manuscript Accepted: January 22, 1986
Published: April 1, 1986
Henri H. Arsenault and Martin Levesque, "Locally adaptive image restoration in the absence of a model for signal-dependent noise," Opt. Lett. 11, 193-195 (1986)