A machine learning model to identify standing water using Keras. API created using Flask.
There are over 1 million deaths annually around the world due to mosquito borne diesases. To combat this issue The Marvelous Mosquito Marauders (Ajay, Amulya, Avni, and Samvrit) created ArboTrack, a crowd-sourced app which allows community members to report geolocation stamped photos of standing waters, a potential mosquito breeding ground, for city officials to treat. Additionally, health officials are able to keep track of potential outbreaks in their area. ArboTrack was selected as one of five National Finalists for the 2020 eCYBERMISSION competition. This convolutional neural network was created with the support of eCYBERMISSION's 2020 STEM-in-Action grant to sort images of standing water from regular water with an 80+% accuracy to filter out spam reports.
You can view their STEM-in-Action grant updates on the eCYBERMISSION Blog: