Figure S3. Three images representative of the image quality scores that were assigned to images included in a Mixed Effect Logistic Regression. The regression tested whether there was a difference in the accuracy of Hotspotter between the Kenyan and Zimbabwean population. The image quality score was included to correct for any potential influence of image quality.

References

Glorot, X., Bordes, A. & Bengio, Y. (2011) Deep Sparse Rectifier Neural Networks. In: Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics. 14 June 2011 JMLR Workshop and Conference Proceedings. pp. 315–323. https://proceedings.mlr.press/v15/glorot11a.html.
Tieleman, T. & Hinton, E. (2012) Lecture 6.5 - rmsprop, COURSERA: Neural networks for machine learning .