Hello sir,
I have read that in the computational processing of the micrographs, the data that we collect during data acquisition is not in the frequency domain, and is converted to the frequency domain via certain algorithms. But I was curious to know that when we import a movie I am sure that information is not in the frequency domain so what cryoSPARC algorithm converts that?
Perhaps I misunderstand your question, but I would say the answer is already in your title: the fast fourier transform is the algorithm that converts the data to frequency domain.
1 Like
yes, I wanted to know, what job does it? Is it Motion correction or CTF estimation?
Everything.
Fast Fourier Transforms are critical to every aspect of modern cryo-EM data processing.
At this point, I would recommend a deeper review of the literature regarding image processing, perhaps starting with:
and
and going down the rabbit hole from there.
4 Likes
Hi @BhawnaMishra!
In addition to the great papers @rbs_sci has recommended, I wonder if you’ve seen Grant Jensen’s cryoEM lectures on YouTube. They’re another friendly resource for getting started thinking about cryoEM, covering sample prep, the microscope, image processing, and some of the math underlying the whole process!
7 Likes
Yes sir, I have started to watch them recently. Thankyou
@BhawnaMishra Because it is the fast Fourier transform, it is rarely useful to persistently store frequency domain images. Further so because such data is both complex (2x storage) and usually zero-padded to a 2x larger size before being transformed (4x storage for 2D, 8x for 3D).
I think the only job types that don’t run FFTs are imports, real-space volume manipution, and metadata handling like exposure sets.