Seminar Series 2020 - Sep 2nd - Imaging Cytometry: Machine Learning

Today's speaker

Peter Rhein (application specialist from Luminex)

02.09.2020 ► 17:00-18:00

Enhancing the Power of Imaging Flow Cytometry by Machine Learning

The Amnis® ImageStream® is a multispectral imaging flow cytometer that combines the statistical power of flow cytometry with imaging content of microscopy in one system. The special features of this technology are based on the fact that it is not only capable of measuring the intensities of fluorescence associated with cells or particles, but it also provides high resolution images of every cell at the same time. These images enable many morphological parameters to be determined, as well as different cell populations to be differentiated accurately and with minimal false positive or negative identification. Thereby, the system enables novel applications that cannot be done with a traditional flow cytometer or a microscope alone.

IDEAS® is the image analysis software included with the Imaging Flow Cytometry systems. The software offers more than 85 algorithms to quantify shape, size, texture, intensity and location in regions of interest defined by masking algorithms. Besides the manual analysis for experienced users, workflows (wizards) for specific applications have been created to reduce the complexity and ramp-up time for new users.

For non-established assays however, more sophisticated algorithms and artificial intelligence are needed. Therefore, a new Machine Learning (ML) module has been added to IDEAS. ML creates an experiment specific feature that will distinguish populations based on user input. It is designed to simplify analysis by allowing users to visually create populations, and enhance discrimination by allowing users to combine multiple fluorochromes and multiple morphologies into a single super-feature.

The workflow of this new module will be shown in a software demonstration to differentiate living and dead cells by their morphology only.

Participate in online webinar (02.09.20 starting at 17:00) using this link => JOIN MEETING