This technology is an algorithm that visually identifies potential breeding grounds (PBGs) of disease vectors to optimize vector-borne disease prevention strategies.
Current methods used to monitor PBGs for vectors, such as mosquitos, rely heavily on human feedback. This can include, for example, user generated phone and email complaints. However, this method of monitoring breeding grounds is inefficient and often imprecise. Furthermore, this method makes it exceedingly difficult to monitor potential breeding grounds over time. Improved methods for monitoring PBDs are needed that address these limitations.
This technology is a system that visually identifies PBGs in images from sources including mobile cameras, CCTVs, satellite images, and thermal camera feeds. The algorithm is first trained on example images to recognize the presence or absence of PBGs such as standing water and garbage piles. It is then able to identify PBGs in source images using Convolutional Neural Networks (CNNs). Geotag information from the image can subsequently be used to notify authorities about the locations of putative PBGs or for navigation to map out “mosquito-free” routes. As such, this technology provides an improved and efficient method for monitoring potential vector breeding grounds with the ultimate goal of reducing PBGs.
Patent Pending
IR CU17077
Licensing Contact: Satish Rao