BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210402T160546Z
LOCATION:Track 10
DTSTART;TZID=America/New_York:20201113T175500
DTEND;TZID=America/New_York:20201113T180000
UID:submissions.supercomputing.org_SC20_sess232_pec242@linklings.com
SUMMARY:Computer Aided Diagnostic Tools for COVID-19 Detection via X-Ray I
 maging
DESCRIPTION:Workshop\n\nComputer Aided Diagnostic Tools for COVID-19 Detec
 tion via X-Ray Imaging\n\nSadre\n\nLightning talk: The goal of this study 
 is to investigate lung geometry and density variations due to tissue and f
 luid alterations due to COVID-19. We approach this problem using two stage
 s. We first have implemented neural networks (NN) as the key image process
 ing method to solve a segmentation problem: detect both lungs. The Python-
 based algorithms explore public datasets as input to U-Net models, a convo
 lutional NN architecture regulated by two main parameters, which influence
  performance: the number of downscaling (and subsequent upscaling) operati
 ons and the number of channels per feature map. Preliminary segmentation r
 esults show that the Dice score reached up to 0.946 using a 5-fold cross v
 alidation on two different datasets, (ieee8023 and NIH).\n\nThe second sta
 ge of our pipeline will involve using this segmentation to isolate the reg
 ions of interest as a preprocessing step for a second neural network that 
 will be used to classify whether a patient has COVID-19 based on their Lun
 g X-ray image. We intend to evaluate the ability of neural networks to cla
 ssify lung images using a large set of 13000+ curated labeled lung X-ray i
 mages using a separated train and test set. We would like to compare the p
 erformances of various popular neural networks on this data set using the 
 raw images as input as well as the segmented images as input in order to q
 uantify the performance benefits of using segmentation as a preprocessing 
 stage for the image analysis pipeline.\n\nTag: Applications, Scientific Co
 mputing, Simulation, State of the Practice, Technology Challenge\n\nRegist
 ration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR