Your mission, should you choose to accept it: track down and identify thousands of nearly microscopic mosquito eggs hiding in plain sight. Your only weapon? A smartphone powered by artificial intelligence.
For a research team project led by FIU’s Andre Costa-da-Silva, it’s a mission possible.
The researchers trained an AI-powered tool that could help mosquito control teams more quickly identify and count mosquito eggs.
The surveillance process begins with ovitraps—small, water-filled containers designed to attract female Aedes mosquitoes seeking sites to lay their eggs. These eggs are deposited on brown paper strips, which specialists have traditionally examined under a microscope. This identification and counting process is both labor-intensive and prone to human error.
Enter AeEggCounter.
The AI-powered tool automatically detects and counts mosquito eggs using images captured with smartphones.
“The concept was to build an application powered by a machine learning model,” stated Costa-da-Silva, an assistant professor of biology and researcher at the Biomolecular Sciences Institute. “This is the pioneer mosquito egg-counting tool designed to analyze authentic field samples through photographs taken on various smartphone devices.”
At the heart of the technology is YOLO—an acronym for “You Only Look Once.” This advanced artificial intelligence algorithm identifies and pinpoints multiple objects within an image in a single pass, enabling rapid and efficient detection. This capability empowers AeEggCounter to accurately detect mosquito eggs directly from smartphone photographs.
The tool is capable of recognizing individual mosquito eggs in both laboratory and field samples because the researchers manually identified and labeled 4,398 individual eggs across 168 smartphone images to train it. Because of their efforts, the AI works even when the tiny eggs are surrounded by small pieces of debris.
The intelligence was in, then it was time to put it to the test.
During the first round, the model correctly identified 88.3 percent of the eggs it flagged and detected 91.7 percent of the eggs actually present in the images.
The next challenge: 76 previously unseen field samples.
Researchers compared the AI’s results with counts made by trained specialists. Although its performance declined, the model’s overall egg counts showed a strong correlation with those of human counters, demonstrating its potential as a tool for mosquito surveillance.
“It doesn’t replace the human,” Costa-da-Silva said. “We still need humans to confirm the results, especially in sensitive areas where it’s important to avoid mistakes.”
The technology is available through ACTT-Study, a free smartphone app for Android and iOS devices.
The study brought together researchers from FIU, the University of South Carolina, Boston University, Baylor College of Medicine, and Texas Children’s Hospital. The findings were published in the Journal of the American Mosquito Control Association.
At FIU, the mission isn’t over.
Matthew DeGennaro, director of the Biomolecular Sciences Institute was a co-investigator on this study is using the app to improve his CDC-funded field studies in Little Havana and East Hialeah.
Melissa Gonzalez and Andre Costa-da-Silva
The technology is also making its way into the classroom.
“Living in Miami was a great aspect to the class because we were able to actually set out ovicups at our homes to screen for eggs,” said Melissa Gonzalez, a second-year biology student. “When we got back to class, we counted them both manually and with the app.”
For Costa-da-Silva, the technology’s potential extends beyond improving mosquito surveillance. It also offers opportunities to train the next generation of scientists in the use of AI-powered tools.
