What Kinds of BioImaging Data Does AI Need? - AskErinDearBeth Episode 32

In this episode, Erin and Beth dive into the questions: What kinds of bioimaging data does AI need? How do we put data together for deep learning models? What kinds of data is useful for deep learning models? How does metadata change that? The field of microscopy doesn't have a perfect answer to these questions yet because "There is no such single thing as a microscope - it is many devices in a trench coat." Beth nicely summarizes the talk with her thoughts on what bio images can do, what we hope they can do in the future, and what work we need to do as a field to get there. #science #openscience #microscopy #tutorial #opensource #biology #imageanalysis #segmentation #cellprofiler ================================================== On Ask Erin/Dear Beth, bioimage analysis experts Beth Cimini and Erin Weisbart, of the Imaging Platform at the Broad Institute of MIT and Harvard, answer your image analysis questions! Whether it’s about tuning or troubleshooting, segmenting or measuring, Beth and Erin have opinions and by golly are they going to share them! Using mostly CellProfiler, but with expertise in many other image analysis softwares such as ImageJ/FIJI, Cellpose, Piximi, and more, no image analysis question is too hard, so ask away! Submit your own questions at broad.io/AskErinDearBeth Images and pipelines we’ve used are available for you to download and play with at https://ciminilab.github.io/AskErin_D...