ZEISS Scanning Electron Microscopes for Industry:

Using established machine learning techniques such as pixel classification or deep learning, even non-expert users can achieve reliable, reproducible segmentation results with ZEISS ZEN Intellesis. Simply load your image, define your classes, label pixels, train your model and perform segmentation. ​

The software only needs to be trained once on a few images to be able to automatically segment batches of hundreds of images. This not only saves time, but also reduces the scope for user-related deviations. All time-consuming segmentation steps on the many similar images are handled by powerful machine learning algorithms.​

ZEISS ZEN Intellesis​
enables particle identification by machine learning and provides higher accuracy for particle identification, learning image segmentation, and object classification.​

Intellesis Object Classification
is used to further classify segmented particles and sort these into their sub-types. This information can then be used to count particles per type. ​

ZEN Intellesis supports easy segmentation of multidimensional images from numerous different imaging sources including widefield, super-resolution, fluorescence, label-free, confocal, light sheet, electron, and X-ray microscopy. ZEN’s evaluation modules then enable automatic report creation and measurement according to industrial standards.​

When it comes to post-segmentation classification by type, ZEN Intellesis takes an innovative approach. Instead of looking at individual pixels as a typical machine learning solution would, its object classification model uses more than 50 measured properties per object to distinguish and classify them automatically. Based on tabulated data, this classification process is much faster than segmentation performed by specifically trained deep neural networks.​

Example for layer thickness:
FIB cross section overlay of CIGS solar cell layers: result from Crossbeam 550 InLens detector (right) and after ZEN Intellesis machine learning segmentation (left).​