What happens when Microscopes gain the power of AI?
Updated: Nov 6
With the power of AI, complex visual procedures get automated. The result is better quality control procedures which ensures that planes stay in the air, food is safe to eat, and medication is properly manufactured.
I have no doubt that everyone reading this article has seen a microscope at least once in their lifetime. It is a tool for examining very small objects. Microscopes are applied in variety of industries in quality control or analytical procedures, as well as in research. Regardless of the application, just like with reading glasses, a microscope is pretty useless without a person using it. In the era of AI – this will change. At Curiosity Labs we developed an AI-enabled microscope that, just like a person, can learn to perform a variety of complex visual assessment tasks. This is a revolutionary breakthrough that can help your lab develop more efficient workflows, and improve quality control through automation.
While you can buy a high-throughput microscope from a variety of manufacturers (BioTek, Nikon, Zeiss and others) – these are only good at rapid image acquisition and have very limited image analysis capabilities. At best, the existing software can perform simple particle (or cell) counting.
AI + Microscope
At Curiosity Labs we realize that image processing is an essential part of the workflow that current microscopes are lacking. The solution is to unite rapid image acquisition capacity of high-throughput microscopes with processing power of AI.
Our microscope, Curiosity Explorer, does just that. The Curiosity Explorer platform comes in two formats – inverted or upright microscope. Both have an automated stage (XYZ movement), can autofocus and take images. What sets it apart from any other microscope is that it has an embedded AI platform. What that means is that your lab can train Curiosity Explorer, just like a technician, to do complex visual microscope assays.
Use cases for an AI microscope
At Curiosity Labs we noticed that numerous manual microscope assays are currently performed by technicians. For example, in Iowa (where our company is based) soil analysis labs receive soil samples, and use microscopes to identify parasitic worms (nematodes). To prove the utility of Curiosity Explorer we trained it to recognize soybean cyst nematode eggs (common soil parasite in the Midwest). The results were incredible! We demonstrated that while people get fatigued rather quickly when using a microscope, but robots don’t. As a result, Curiosity Explorer demonstrated >95% accuracy at recognizing soybean parasites, while human accuracy decreases to 80% during the onset of fatigue. In addition, Curiosity Explorer is up to 15 times faster than a person. The big picture is that Curiosity Explorer can be trained to recognize any soil organism or particle with accuracy and speed greater than that of a person.
Implementation of AI into microscopes is a significant breakthrough that impacts applications outside of agriculture. For example, we are currently exploring the possibility of training our AI platform to recognize anomalies in thermal spray coating procedures, which is a common procedure in aircraft manufacturing (at companies like Boeing, Rolls Royce, GE). The automation of this procedure will significantly improve the efficacy of quality assurance of critical parts that keep our planes in the air.
Another application for Curiosity Explorer is in automating quality control of food stocks that are used to make a variety of products that we eat every day (at companies like Nestle and others). Our AI platform can detect foreign particles in food with much higher success than people.
AI + You
What about you? Without realizing it, even an average reader will benefit from AI microscopes. This technology ensures that our planes stay in the air, the food is safe to eat, and our medication is safe.
At Curiosity Labs we are eager to find new applications for our AI-enabled microscope, Curiosity Explorer. If your lab currently uses microscopes for quality control – reach out to our team to explore how Curiosity Explorer can help you.
President of Curiosity Labs.