Florida researchers develop AI-powered trap to catch dengue-carrying mosquitoes with over 98% accuracy

University of Florida researchers have developed an AI-powered mosquito trap that could help curb dengue transmission by detecting disease-carrying mosquito species. The release of the new technology comes at a time when Florida is battling through one of the largest dengue outbreaks, with over 700 cases of the mosquito-borne disease as of October 22.

Created by Associate Professor Ryan Carney and Professor Sriram Chellappan and commercialized by USF startup Digitomy, this AI-powered smart trap “has achieved more than 98% accuracy in identifying” dengue-causing mosquito species during laboratory testing, the Tampa-based university said in a media release on October 23.

“The mosquito is the deadliest animal on the planet. While there are 3,500 mosquito species, it's primarily only one, Aedes aegypti, that is spreading dengue in Florida. We can use the data from our smart traps to construct a habitat model, as well as to monitor the species' presence and abundance in real time, to strategically mitigate its spread,” said Carney in a statement.

How this AI-powered mosquito trap works

The trap, when hung from a low tree branch or placed on the ground, can attract mosquitoes with a light bait or an in-built puling fan. A tickty pad was placed inside to hold them while an AI camera scans them for identification, matching them against a vast database of mosquito images.

Also Read: A Soviet physicist put his head inside a ginormous particle accelerator in 1978; a proton beam 600 times the radiation dose a human could survive pierced his skull, peeled off his skin and paralyzed half his face, but he survived

“Traditional mosquito surveillance requires a long manual approach by specialists, which can take two to three days under a microscope,” Chellappan said. “The AI in our trap speeds up the process and can lead to more targeted disease interventions, while still being human-led,” he added.

It takes less energy than sharing a single Instagram post

As per the researchers, this is also highly energy efficient; it can identify an entire sticky pad of mosquitoes with less energy compared to sharing a single Instagram post. Other than identifying mosquitoes, Digitomy could help public health officials pinpoint where the disease-carrying species are concentrated.

The start-up is also developing a cloud-based education platform that the general public can access. There, they can examine mosquito images, identify species, and eventually receive AI-generated feedback to improve their identification skills.

“Education is our secondary line of defense,” Carney said in the statement. “While our smart traps provide the macro-level data, the public should practice the three D's: Drain standing water, Dress to cover skin and wear DEET. Pairing automated AI data with mosquito control and engaged communities can help us conquer these outbreaks.”

Diana Shapiro, chief operations officer of Digitomy and operator for several USF ventures in USF Technology Transfer's Launch Program, said, “mosquito-control programs wait days to learn what is in their traps,” adding, “our technology is built to tell them the same day, and to reach the places with the least to spend and the most at stake.”