Computer Vision for Drought Resilience
deep learning for better index insurance
ANNOUNCEMENT: I’ve launched a benchmark competition in partnership with Weights and Biases. If you’re a computer vision expert or are interested in getting started in computer vision and/or deep learning, please give it a shot! The best model will be used to create a better insurance product for Kenyan pastoralists.
Droughts are the greatest risk facing pastoralists and their families in Northern Kenya. When droughts strike, the livestock that are the basis of their livelihoods and food supply starve and stop producing milk. Index insurance can make droughts much less damaging by providing pastoralists with money to weather droughts. The more accurate the index, the better the insurance. This project is an experiment to test the ability of deep learning and computer vision to
This project was part of the Computer Vision for Global Challenges Workshop at the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) in Long Beach, California.