Image quality improves significantly as the number of computer processing iterations increases from 0 (left) to 500 (right) (Image: The American Physical Society)
For years physicists have grappled with a kind of double vision that has made using holography difficult. The process results in an out-of-focus second image being superimposed on the main one, which can seriously degrade the result. "The twin image problem has existed since holography was conceived. People have always worried about it," says Hans-Werner Fink, a physicist at the University of Zurich in Switzerland.
Physicists have devised various optical techniques for removing the twin, but they do not always work for light of much shorter wavelengths such as X-rays.
Now Fink and his colleague Tatiana Latychevskaia have solved the problem in a way that should work, regardless of the source of illumination.
Latychevskaia says she realised how to do it after noticing that the blurry superimposed twin makes some areas of the light field brighter than would be possible were there a single image alone.
So she designed a computer program that identifies these regions, replaces them with a more realistic light level, and than calculates how this would affect the light field that created them.
The modified light field is then used to create a new image and the process begins again. Repeating this process many times removes the twin image entirely, dramatically sharpening the result.
"Although tested so far only in the optical region, there is no obvious reason why the authors' method should not also work with X-rays and electron waves," he told New Scientist. That would be important for using the technique on a microscopic scale.