Deep Learning Segmentation of Satellite Imagery Identifies Aquatic Vegetation Associated with Snail Intermediate Hosts of Schistosomiasis in Senegal, Africa

Schistosomiasis is a debilitating parasitic disease of poverty that affects more than 200 million people worldwide, mostly in sub-Saharan Africa, and clearly associated with the construction dams water resource management infrastructure tropical subtropical areas. Changes to hydrology salinity linked development may create conditions favorable aquatic vegetation suitable habitat for intermediate snail hosts schistosome parasites. With thousands small large reservoirs, irrigation canals, developed or under it crucial accurately assess spatial distribution high-risk environments are freshwater schistosomiasis rapidly changing ecosystems. Yet, standard techniques monitoring snails labor-intensive, time-consuming, provide information limited areas can be manually sampled. Consequently, low-income countries where control most needed, there formidable challenges identifying potential transmission hotspots targeted medical environmental interventions. In this study, we new framework map across scales Senegal River Basin by integrating satellite data, high-definition, low-cost drone imagery, an artificial intelligence (AI)-powered computer vision technique called semantic segmentation. A deep learning model (U-Net) was built automatically analyze high-resolution imagery produce segmentation maps vegetation, fast robust generalized prediction proved accurate commonly used random forest approach. Accurate up-to-date knowledge at highest risk increase effectiveness interventions targeting disease-carrying snails. deployment framework, local governments health actors might better target when they needed integrated effort reach goal elimination.