Background Blob in Full Frame Motion Correction Exported .cs Files


I’ve noticed that the exported .cs files for full frame motion correction include fields that refer to a ‘background blob’. I assume this is meant to be an estimate of the background shade of a given image, but I was interested in knowing how it is calculated and used. I was puzzled, because, for one example I looked at, the dimensions of the .mrc file pointed to by ‘background_blob/path’ were square despite the original tiff image itself not being square (The dimensions were also a power of 2, which I suspect is relevant to how it is calculated in some way). I was wondering if anyone had some insight on what the purpose of the background blob is and what it represents.


Information in the background_blob MRC-format files is used to correct for variable ice thickness across a microscopic frame. The power-of-2 dimensions facilitate speedier computations, which include Fourier transformations and low-pass filtering. The potential discrepancy with the original aspect ratio is due to padding, which helps avoid artifacts during Fourier transformation and is applied at variable widths to achieve a square aspect ratio. Were it not for the padding, background_blob frames would match the aspect ratio of the underlying movie frames.