MozWatch is an app that differentiates between 3 different genera of mosquitoes when users upload a picture of a mosquito found in their home. Using a machine learning model, the app classifies the mosquito as being Aedes, Anopheles, Culex, or not a mosquito.
Knowing the genus of the mosquito aids in understanding where it may be breeding in a home, thereby allowing users to address the issue from its root. Further, understanding when and where mosquitoes may bite allows users to protect themselves; for instance, by wearing mosquito repellant during certain times in the day.
The machine learning model is an image classification model made with Apple's CreateML. The model has an accuracy of 88%, and was trained on 4000+ images of mosquitoes from the dataset from Pise, et al.
(Pise, Reshma; PATIL, Kailas; Laad, Meena; Pise, Neeraj (2022), “Dataset of Vector Mosquito Images ”, Mendeley Data, V4, doi: 10.17632/88s6fvgg2p.4 )
Yes, all information in the app has been verified and edited by an industry professional from the National Environment Agency of Singapore.