Researchers have ventured into the developing of an artificial intelligence program which can assist in providing information about species-level identification of microscopic organisms in the ocean. They are yet to instill the program into a robotic system which will enable the understanding of the ocean from the past to present.
The program invented has so far been able to identify six species of the forams organisms also known as foraminifera that are mostly found in the Earth’s ocean for the last 100 million years. Foraminiferas are neither plants nor animals also known as protists such that when they die, they leave tiny shells behind. These shells are not more than a millimeter wide.
The shells are used by scientists to get information about the characteristics of the oceans as it was when they were still alive. A good example is whereby a number of forams exist in different types of chemical measurements and ocean environments. Studying the shells can give researchers all the details from the temperature and the chemistry of the ocean when the shell was formed.
The process of evaluating foram shells and their fossils can be very time consuming and tedious. To make this easier, a team of researchers and expertise form the fields of robotics and paleoceanography are working to turning it to an automated process.
According to Edgar Lobaton, an associate professor of electrical and computer engineering from the University of North Carolina, the AI can accurately identify the forams 80 percent of the time which is way better than trained humans.
The system currently works by placing a foram under a microscope, and a LED ring shines on the foram, and 16 directions at a time while taking images of the foram. The 16 images are put together to form geometric information, and the AI uses this to identify the species of the forum.