~/talk

The Design and Evaluation of a Mobile System for Rapid Diagnostic Test Interpretation

Talkat UbiComp 2021 · 2021-09-21..09-26

speakers

name:Chunjong Park

links:none on file

links:none on file

name:Libby Rose Lavitt

links:none on file

name:Vincent Karuri

links:none on file

name:Shiven Bhatt

links:none on file

name:Peter Lubell-Doughtie

links:none on file

name:Anuraj H. Shankar

links:none on file

name:Leonard Ndwiga

links:none on file

name:Victor Osoti

links:none on file

name:Juliana K. Wambua

links:none on file

name:Philip Bejon

links:none on file

name:Lynette Isabella Ochola-Oyier

links:none on file

name:Monique Chilver

links:none on file

name:Nigel Stocks

links:none on file

name:Victoria Lyon

links:none on file

name:Barry R. Lutz

links:none on file

name:Matthew Thompson

links:none on file

name:Alex Mariakakis

links:none on file

name:Shwetak N. Patel

links:none on file

resources

video:https://youtu.be/Kr9w3j89FIc

slides:none on file

summary

A smartphone app called RDTScan guides users with real-time feedback (brightness, blur, position) and captures RDT images without hardware attachments, using Laplacian sharpness checks, histogram/CLAHE on the HLS L channel, trough detection for lines, and SIFT feature matching to align templates. The system configures new tests from one template image plus metadata and shows a limit-of-detection comparable or better than experts in lab tests; field studies report 97.5% (Australia flu) and 96.3% (Kenya malaria) accuracy.

AI-written summary · about