MegaLUT is a simple and fast method to correct ellipticity measurements of galaxies from the distortion by the instrumental and atmospheric point spread function (PSF), in view of weak lensing shear measurements.
The method performs a classification of galaxies and associated PSFs according to measured shape parameters, and builds a lookup table of ellipticity corrections by supervised learning. We have applied this new method to the GREAT10 image analysis challenge, and present in our paper a refined solution that obtains the competitive quality factor of Q = 104, without any power spectrum denoising or training. Of particular interest is the efficiency of the method, with a processing time below 3 ms per galaxy on an ordinary CPU.
Here’s the idea (click for larger image) :