State of the Art in denoising images with high PSNR and without access to original sources?

Hello,

by writing a report about denoising images without access to the
original  I want to know what were the latest cognitions in denoising
images with higher signal-to-noise ratio.

I found various papers, pE. from Peters II/Alan (A New Algorithm for
Image Noise Reduction using Mathematical Morphology,1995), Donoho (
De-noising by soft-thresholding, 1995), Hsia/Chan (A FAST EFFICIENT
RESTORATION ALGORITHM FOR HIGH-NOISE-RATIO IMAGE USING FEATURE ADAPTIVE
APPROACH, 2002) and I analyzed the stuff around transcode-/virtualdub
filters.

IMHO the most popular methods were open/closing and weighted mean/median
filtering. I am right? What were also important papers and/or
technologies to handle with noisy images without access to the original
image (in that way, knowledge and channel estimation could be modelled)?

Which criteria is most helpful to build an ergonomic, clean image (from
the humans eye view) from a highly noised one?

Which methods have a good balance between quality and calctime
consumption, which were realtime operable?

Thanks,

Bye Andreas
-- 


--