BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//euroscipy-2026//talk//FMNMDA
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-euroscipy-2026-FMNMDA@pretalx.com
DTSTART;TZID=CET:20260721T113000
DTEND;TZID=CET:20260721T120000
DESCRIPTION:The transition toward Organ-on-Chip (OoC) and Engineered Heart 
 Tissues (EHTs) in pharmacological research necessitates automated tools fo
 r longitudinal cell monitoring. Manual tracking of cardiac cycles is labor
 -intensive and impractical for long-duration studies\, highlighting the ne
 ed for high-throughput segmentation. This study proposes a specialized sha
 llow 2D U-Net architecture utilizing Focal Loss to effectively manage clas
 s imbalances between cellular structures and the background in spatial rec
 ordings.\n\nThe model was trained on a dataset of 1\,240 images and valida
 ted against 100 test samples\, achieving a Dice score of 0.9152 and a mean
  Intersection over Union (mIoU) of 0.8486 relative to manual ground truths
 . To contextualize these results\, the proposed network was benchmarked ag
 ainst MedSAM\, a state-of-the-art foundation model for medical imaging.\n\
 nWhile MedSAM offers broad generalization\, our lightweight U-Net demonstr
 ates superior computational efficiency and localized precision for the hig
 h-frame-rate requirements of cardiac cycle analysis. The findings suggest 
 that a task-specific\, loss-optimized architecture provides a high-perform
 ance\, resource-efficient alternative to large-scale foundation models\, f
 acilitating objective and scalable analysis of single-cell behavior in ani
 mal-free drug testing environments.
DTSTAMP:20260603T195631Z
LOCATION:Room 1.19 (Ground Floor\, Shannon)
SUMMARY:Comparative Analysis of Focal Loss-Optimized Shallow Convolutional 
 Neural Network and MedSAM for Precise EHT Segmentation in Dynamic Spatial 
 Recordings - Maciej Szymkowski
URL:https://pretalx.com/euroscipy-2026/talk/FMNMDA/
END:VEVENT
END:VCALENDAR