Currently, there is no open-source and low-cost system to detect low concentrations of bacteria in water samples. The undergraduate researchers at Fort Lewis College developed an optical system to quantify single bacterium from environmental water samples (top).

To improve signal detection in their system, they developed a neural network, a method in artificial intelligence. They trained this neural network to recognize the difference between single droplets of bacteria, multiple droplets, and sample noise.

The system successfully detected samples of various sized fluorescent particles with 0% error. The next steps will be to implement this system on the biological samples.