Mic junk detector showing offset masks

Exposure Curation
1

Hi,

I have encountered an odd bug (?) with Mic. Junk Detector. When I run with input from a Patch Motion job run with output Fcrop=1/2, the results are as expected:

J513_example_4_of_20_j2motioncorrected008221330530316764804_23dec04c_nbc11_black4_00001gr_00025sq_v01_00002hln_v01_00005ennframes_patch_aligned_doseweightedmrc
J513_example_4_of_20_j2motioncorrected008221330530316764804_23dec04c_nbc11_black4_00001gr_00025sq_v01_00002hln_v01_00005ennframes_patch_aligned_doseweightedmrc427×303 120 KB

But when I run with input from an identical Patch Motion job, except that it was run with Fcrop=1, the junk masks are offset:

J512_example_4_of_20_j324motioncorrected008221330530316764804_23dec04c_nbc11_black4_00001gr_00025sq_v01_00002hln_v01_00005ennframes_patch_aligned_doseweightedmrc
J512_example_4_of_20_j324motioncorrected008221330530316764804_23dec04c_nbc11_black4_00001gr_00025sq_v01_00002hln_v01_00005ennframes_patch_aligned_doseweightedmrc506×359 171 KB

The raw pixel size was 0.5135 Å. Both Patch Motion jobs were run in v4.7.

I suspect it is something to do with the relatively small pixel size, as I do not observe this for datasets with a raw pixel size ~> 1 Å. I tried on another high mag dataset, EMPIAR-11918 (Apix=0.531 Å), with the same results:

Fcrop=1/2:

J265_example_6_of_20_j264motioncorrected001431403632209060749_foilhole_8466724_data_8467440_8467442_20230209_185527_fractions_patch_aligned_doseweightedmrc
J265_example_6_of_20_j264motioncorrected001431403632209060749_foilhole_8466724_data_8467440_8467442_20230209_185527_fractions_patch_aligned_doseweightedmrc427×303 127 KB

Fcrop=1:

J263_example_6_of_20_j262motioncorrected001431403632209060749_foilhole_8466724_data_8467440_8467442_20230209_185527_fractions_patch_aligned_doseweightedmrc
J263_example_6_of_20_j262motioncorrected001431403632209060749_foilhole_8466724_data_8467440_8467442_20230209_185527_fractions_patch_aligned_doseweightedmrc506×359 176 KB

The displacement seems to be more substantial the further away the junk area is from the origin (lower left-hand corner).

Cheers
Oli

2

Thanks @olibclarke for posting the example images. Were these images screenshots from a specific job type: Micrograph Junk Detector? Curate Exposures?

3

Directly from the log file of Junk Detector IIRC

EDIT:

Confirmed @wtempel - it was from the Junk Detector log. But same effect is visible in Exposure Curation, so it is not only an issue with that log:

image
image1552×1122 257 KB

image
image1536×1106 256 KB

4

@olibclarke This has been fixed in the recent release (4.7.1). Enjoy.

1 Like