Lianrui Zuo
Atlas
2000

Transformation and interpolation

An image has values only at its pixels. How do we move it, when every moved pixel lands between them?

Primary source P. Thévenaz, T. Blu and M. Unser, “Interpolation revisited,” IEEE Transactions on Medical Imaging, vol. 19, no. 7, pp. 739–758, 2000. doi:10.1109/42.875199

An image has values only at its pixels, so turning it means guessing the values in between. Choose a way to guess: take the nearest pixel, join neighbouring pixels by straight lines, or fit a smooth curve through them. Then turn the image by 30° twelve times, a full turn, which should bring it back exactly. The number is how far the result is from the original: the guesses pile up, and the three ways pile up differently.

The full lecture is open to students in ECE 6357. Enter the course password to read it.