BDemps
commented
7 years ago Hi Karl, This description seems like a good one for the general considerations.
On your issues to consider: (1) I guess exact numbers won't be absolutely necessary as long as we are in the ballpark and don't exclude real data, right? (2) Can we just crop based on the unrotated data to get a "close enough" measure? (3) Yes, true. (4) We probably want to have an intersection crop, but again it can be a ballpark estimate based on unrotated data, I would think.
On Wed, Oct 4, 2017 at 2:07 PM, Karl Czajkowski notifications@github.com wrote:
This needs more thought...
In some cases, a region identified by one crop may need a secondary QA process where parts of the image boundary are considered below standard and a narrower crop introduced.
Because segment classification may already be underway at this point or completed in the past, we don't want to redefine the actual crop region used for the classification tasks. Instead, we may want to introduce a secondary crop which can be used by analysts to discard some puncta which fall outside the secondary region, which MUST be a proper subset of the first region.
Issues to consider:
- Timepoint 1 pixel coordinates map directly to micron coordinates with simple scaling.
- Timepoint 2 pixel coordinates are not only scaled but also potentially rotated and displaced by image pair registration.
- Ideally, we only want analysts to have to deal with a simple axial range restrictions to discard points, without having to do general 3D geometry and plane clipping.
- Do we care just about clipping individual region timepoints, or do we need to come up with an intersection crop that considers only the common volume inside both regions or both regions with secondary crops?
@carlkesselman https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_carlkesselman&d=DwMCaQ&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=Lssj6JKak_DaRnExGYcKnQ&m=w-UwMV2l6RaLFaZGBb5riaqrm0ZGXIvnGxflRTLfAY8&s=DGCeCPtc-lOhX5n5bl8ISv9qKcJfhWEhPSI7uP9nInA&e= @BDemps https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_bdemps&d=DwMCaQ&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=Lssj6JKak_DaRnExGYcKnQ&m=w-UwMV2l6RaLFaZGBb5riaqrm0ZGXIvnGxflRTLfAY8&s=DG0SG22DbzKY0KAZmC50o4ibngk1ARN8KC2N35_MMAo&e=
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_informatics-2Disi-2Dedu_synspy_issues_48&d=DwMCaQ&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=Lssj6JKak_DaRnExGYcKnQ&m=w-UwMV2l6RaLFaZGBb5riaqrm0ZGXIvnGxflRTLfAY8&s=Nk4RJ2zqYtl8QyWSdNL_CpNe4ZSjJ0PhlSkICWEsk3o&e=, or mute the thread https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_AbQSt3a2G-5FSlTxjkAc-2DJoQwr4zPdxNkcks5so-5FOPgaJpZM4PuPS3&d=DwMCaQ&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=Lssj6JKak_DaRnExGYcKnQ&m=w-UwMV2l6RaLFaZGBb5riaqrm0ZGXIvnGxflRTLfAY8&s=0jiglDN3ypmBIW_q_jynVJ6t3bd9dCluKah5tQUeLNI&e= .
-- William Dempsey, Ph.D. Postdoctoral Scholar Molecular and Computational Biology Dana and David Dornsife College of Letters, Arts and Sciences University of Southern California wdempsey@usc.edu Lab #: 213-821-1818
This needs more thought...
In some cases, a region identified by one crop may need a secondary QA process where parts of the image boundary are considered below standard and a narrower crop introduced.
Because segment classification may already be underway at this point or completed in the past, we don't want to redefine the actual crop region used for the classification tasks. Instead, we may want to introduce a secondary crop which can be used by analysts to discard some puncta which fall outside the secondary region, which MUST be a proper subset of the first region.
Issues to consider:
@carlkesselman @BDemps