Engineers have created a microscope that adapts its lighting angles, colors, and patterns while teaching itself the best settings for a given diagnostic task.
The microscope developed a lighting pattern and classification system simultaneously in the engineering team’s study, allowing it to quickly identify red blood cells infected by the malaria parasite more accurately than trained physicians or other machine learning approaches.
A standard microscope illuminates a sample with the same amount of light coming from all directions, and that lighting has been optimized over hundreds of years for human eyes. However, computers can see things that humans cannot, so he and his team’s hardware redesign includes a diverse range of lighting options, allowing the microscope to optimize its illumination.
Instead of diffusing white light from below to illuminate the slide evenly, the engineers created a bowl-shaped light source with LEDs embedded throughout its surface. This enables samples to be illuminated from various angles up to nearly 90 degrees with different colors, casting shadows and highlighting different features of the sample depending on the LED pattern used.
The researchers then fed hundreds of samples of malaria-infected red blood cells prepared as thin smears on a microscope slide, with the cell bodies remaining whole and ideally spread out in a single layer. The microscope learned which features of the sample were most important for diagnosing malaria and how to highlight those features using a type of machine learning algorithm known as a convolutional neural network.
Related Content: Early-Stage Lung Cancer Detection