Topographical Computer Vision Algorithms for rapid, low-cost Hematological diagnostics and Parasite detection through Random Forests classification and van Leeuwenhoek-type ImagingTanay Tandon, USA Finalist, 16-18
This research develops a low-cost device for automated, portable blood diagnostics and parasite detection on a smartphone system. Rural regions lack access to expensive in-lab diagnostic equipment and trained pathologists for disease detection, leading to millions of undiagnosed cases and high parasite mortality rates. The developed artificial intelligence model utilizes machine learning to interface with a lens system attached to a phone camera, automatically identifying parasites and diseases in blood with 0.73-0.85 accuracy. The vision model was trained on CDC blood-smear datasets, and the device was tested on smears such as Chagas disease, Sickle Cell Anemia, and Whole Blood samples.