Abstract: Vector surveillance is critical to integrated vector management. However, identification is often time consuming and requires significant expertise. For mosquitoes, some vector control organizations have significant regional expertise in identification of the local species. However the task is time consuming and arduous, requiring experts to sift through the thousands of specimens captured by a fleet of traps multiple times a week. With funding constraints, many organizations delegate this task to seasonal workers, resulting in a high training burden and variable accuracy and speed. Furthermore, many organizations are beginning to expand operations to include tick surveillance. Tick identification requires its own specialized training, creating a barrier for many vector surveillance organizations who aim to create or expand active tick surveillance programs. To help address these gaps, Vectech is developing IDX, a versatile tool for the imaging and identification of small arthropods, with an initial focus on critical vectors for disease such as mosquitoes and ticks. To date, development of tick species identification capabilities have achieved 98.5±0.3% macro-averaged F1 score on IDX image datasets including 12 species, nine of which are commonly found in the US, including Rhipicephalus sanguineus, Ixodes pacificus, Ix scapularis, Dermacentor variabilis, Haemaphysalis longicornis, and Amblyomma americanum. Development of mosquito species identification has continued to include 23 species, achieving 94% macro-averaged F1 score including Aedes aegypti, Ae albopictus, Anopheles albimanus, An gambiae sl, An stephensi sl, and Culex pipiens sl. IDX identification algorithms are deep learning based, using convolutional neural networks, which require high data volumes. To this end, 12 IDX devices have been distributed to early partners who aid in gathering image data and algorithm assessment for both ticks and mosquitoes.
