ZEISS

Microscopy is the umbrella term for all methods used to observe tiny objects. ZEISS has been setting standards in this field since the mid-19th century, and no fewer than 30 Nobel Prize winners have used ZEISS optical solutions. Today, as a leading manufacturer of microscopy systems, the company provides groundbreaking solutions and services, including both hardware and software. To streamline the digital analysis of microscopy images, ZEISS now offers researchers and businesses the ZEISS arivis Cloud platform, featuring deep learning training tools for image segmentation. The solution is built on Microsoft Azure.

Pick any industry and chances are it relies in some way on microscopy and image analysis—from life sciences to materials sciences to earth sciences. Professionals in these disciplines must be able to get a close look at what are often incredibly complex structures. Materials scientists are interested in looking at the microstructure of different materials, while bioscientists want to examine human cells. While many aspects of image analysis are highly automated, one crucial step lags behind: image segmentation. However, this is undergoing transformation.

The challenge: Automating the manual image segmentation

“Segmentation involves dividing an image into different regions that are crucial for subsequent analysis and classification,” says Dr. Sandra Lemke, Product Owner for the AI toolkit on ZEISS arivis Cloud at ZEISS Microscopy. “This step is absolutely essential as it provides researchers with the foundation for the precision of the analyses that follow. It’s how a life sciences researcher defines a cell or a nucleus. Without the help of Artificial Intelligence, they’d even have to manually draw the structures for some images pixel by pixel.”