Modified Mosquito Programs’ Surveillance Needs and An Image-Based Identification Tool to Address Them

Effective mosquito surveillance and control relies on rapid accurate identification of vectors confounding sympatric species. As adoption modified (MM) techniques has increased, the value monitoring success interventions gained recognition pushed field away from traditional ‘spray pray’ approaches. Field evaluation MM that target specific species require massive volumes data involving species-level identifications. However, methods remain time labor-intensive, requiring highly trained, experienced personnel. Health districts often lack resources needed to collect essential data, conventional entomological involves a significant learning curve produce consistent high accuracy data. These needs led us develop MosID: device allows for high-accuracy enhance capability capacity programs. The features high-resolution optics enables batch image capture specimens using computer vision. While development is ongoing, we share an update key metrics MosID system. algorithm, tested internally across 16 species, achieved 98.4 ± 0.6% % macro F1-score dataset known unknown used in training, reserved testing (species, respectively: 12, 1302; 603; 7, 222). Preliminary user showed were processed with at rate ranging 181-600 per hour. We also discuss other within technical scope, such as sex fluorescence detection, may further support