BEGIN:VCALENDAR
VERSION:2.0
PRODID:icalendar-ruby
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VTIMEZONE
TZID:Europe/Vienna
BEGIN:DAYLIGHT
DTSTART:20260329T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20251026T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260424T143149Z
UID:1769158800@ist.ac.at
DTSTART:20260123T100000
DTEND:20260123T110000
DESCRIPTION:Speaker: Marten Chaillet\nhosted by Alicia Michael\nAbstract: C
 ryo-electron tomography (cryo-ET) is used to visualize complex cellular en
 vironments at macromolecular resolution. However\, due to limitations of t
 he microscope\, computational image alignment is crucial for data interpre
 tation. Existing reference-free alignment algorithms aim to maximize simil
 arity between adjacent tilt images for improved registration. This often p
 roves ineffective due to limited information overlap between images and in
 accurate assumptions about the sample. Meanwhile\, human experts can easil
 y recognize misalignment. We introduce a new machine learning-based approa
 ch for training similar intuition and using it to improve alignment. MissA
 lignment trains a convolutional neural network to score the alignment accu
 racy using a contrastive loss metric that doesn’t require well-aligned g
 round truth. Back-propagation from this score is then used to optimize ind
 ividual image alignment parameters. Our method demonstrates significantly 
 improved alignment compared to existing techniques\, leading to superior p
 erformance in all downstream analysis tasks. This advancement substantiall
 y enhances the robustness of cryo-ET data processing\, making the techniqu
 e applicable to a broader range of samples.
LOCATION:Office Bldg West / Ground floor / Heinzel Seminar Room\, ISTA
ORGANIZER:alicia.michael@ista.ac.at
SUMMARY:Marten Chaillet: MissAlignment learns to straighten out cryo-ET til
 t series
URL:https://talks-calendar.ista.ac.at/events/6230
END:VEVENT
END:VCALENDAR