Archive notice: This article was originally published on August 1, 2005. Links and embedded videos are preserved as part of the historical record.
In the half-day course “Computational Photography”, several image processing methods were presented that make it possible to get the maximum image quality out of conventional digital photos.
With the help of the gradients developed by Ramesh Raskar and Jack Tumblin at the Mitsubishi Electric Research Labs (MERL), it is possible, for example, by photographing a scene twice to change or even shift the depth-of-field range afterwards on the computer.
It was also shown how an underexposed shot can be improved with the help of a second shot with flash. Both images has very poor quality when viewed on their own: If they are calculated together, however, you get a crystal-clear, optimally illuminated image. The gradients also make it possible, for example, to remove fog from a digital photo without affecting the background. All in all, a well-attended lecture that had it in itself.
Further Links:
- Erg
