**5. Conclusions**

In conclusion, the reliance on manual scoring of mosquito egg development to determine the impact of PPF on the fertility of *Anopheles* mosquitos requires a level of expertise and experience that is not always available. This paper shows these limitations can be overcome using a ResNet-50 CNN model that automates the classification of egg development as a measure of fertility status. Such a model is fast, can achieve an acceptable level of precision, is in a robust format, and has the potential to be easily distributed in a practical, accessible, and freely available manner. Furthermore, this approach is applicable to scoring the fertility of mosquitoes exposed to any PPF-treated tool or similarly acting insecticide or insect growth regulator, which causes the same impact on ovary morphology, as well as applicable during bioassays performed to measure efficacy of these tools and in resistance monitoring.

**Author Contributions:** Conceptualisation, A.S. and R.S.L.; methodology, analysis, and writing (draft preparation), M.T.F.; materials, R.S.L., J.F., N.S.M., N.P. and C.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Bill and Melinda Gates Foundation, gran<sup>t</sup> number INV-004350, through Innovation to Impact (I2I) at the Liverpool School of Tropical Medicine. M.F. is funded by the Medical Research Council of the UK, gran<sup>t</sup> number MR/P027873/1, through the Global Challenges Research Fund.

**Institutional Review Board Statement:** Not applicable.

**Data Availability Statement:** All code and model weights used for analysis, modelling, and visualisation are available in a GitLab repository at https://gitlab.com/MTFowler/lstm\_ovaryclassify (last accessed on 3 December 2021). As the images in this study are still being used in ongoing research, they will only be available upon reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest.
